Home

Report on the LRAIC Model and User Guide

1. Layer 2 node Layer 2 node L uil iti on app 1000 sites y y ayer 3 node on app 100 sites Layer 3 node Layer 2 ring Layer 3 node IDSLAM DSLAM Switch Switch Switch _ Switch Switch IDSLAM Router Router Layer 2 node L Layer 2 ri Layer 3 ri Switch ayer 2 ring ayer 3 ring DSE AM IP MPLS backbone Layer 2 node Router Router Switch Switch Switch Switch Switch DSLAM DSLAM DSLAM Switch Layer 2 node Layer 3 node Layer 3 node DSLAM Layer 2 ring 450 nodes in bus Layer 2 node 7 Doe es Layer 2 node Layer 2 node In the 2006 model revision it was assumed that OAOs would connect via a Layer 2 switch Following discussions with NITA it was agreed that three specific BSA products would be addressed all of which would assume Ethernet based backhaul gt BSA at the Street Cabinet gt BSA at the Layer 2 Ethernet switch location approximately 1000 sites on the Layer 2 rings with around a further 450 sites feeding off on of those ring locations gt BSA at the Layer 3 IP Router location approximately 100 sites on the Layer 3 rings The network modelled is essentially that which would exist following the completion of the Street Cabinet upgrade programme Additional inputs were th
2. DSLAM lines DSLAM planning persons 1000 00 DSLAMs Access infrastructure management persons Core nodes Access infrastructure maintenance persons NTPs Access infrastructure planning persons Core nodes others to be added if needed persons One Off organic growth persons IMS installation persons IMS Core infrastructure installation persons 100 100 600 00 km Core trench Core DWDM installation persons 500 00 DWDM IP Network installation persons 1000 14 00 Core nodes DSLAM installation persons o wwo aoo 400 00 DSLAM Access infrastructure installation persons o 1000 400 35 00 Core nodes others to be added if needed persons a aa Source Consolidation model I_FA_Costs 3 4 2 Allocation of costs Three types of operational costs are to be allocated gt Network related costs which are directly related to operating the network e g network management equipment maintenance or access installation gt Interconnection related costs which are directly related to the commercial operation of interconnection e g subscriber support and interconnection accounting systems gt Non network related costs or other operational costs which are necessary for running an organisation For example HR costs Board of Directors costs legal assistance etc 3 4 2 1 Allocation of network related costs The division into FA areas may not be used directly for allocating the network related costs Therefore a m
3. Date 30 11 2009 Page 23 To consider this extra indirect cost derived from other personnel s time consumption a 25 percent mark up has been applied to all three wage categories In version 2 4 of the model the list of functional areas was clearly based on an underlying circuit switched network and thus for the revised model the areas needed to be adapted to reflect an underlying packet switched network Furthermore the staffing levels in version 2 4 were fixed for each functional area which is now felt to be no longer appropriate given the more rapidly changing nature of both the quantities and types of equipment in IP based networks In addition to updating the list of functional areas to reflect the IP based network modelled the fixed quantities of staffing have been replaced with variables that can flex to some degree with underlying demand From version 2 5 for each functional area it is now possible to specify a minimum level of staffing together with an additional number of staff per so many of a given driver The list of drivers currently in the model is illustrated in Table 3 4 but could easily be adapted should more suitable drivers be identified later in consultation with the industry Table 3 4 List of Opex Drivers 1 1 Opex Drivers Quantity IMS 13 Core nodes 1 838 km Core trench 14 454 DSLAMs 4 640 DSLAM lines POTS and xDSL combined 2 582 065 NTPs 2 917 626 DWDM 46 Source Consolidation mode
4. PSTN Public Switched Telecommunications Network RTP Real Time Protocol RTU Right to Use licence SBC Session border controller SDH Synchronous Digital Hierarchy SDP Secondary Distribution Point Date 30 11 2009 Page 129 SDSL Symmetrical Digital Subscriber Line SMP Significant Market Power st rk markedsposition Subs Subscribers TDM Time Division Multiplexing TX Transmission VAD Voice Activity Detection VDSL Very High Bit Rate Digital Subscriber Line VoIP Voice over IP xDSL x Digital Subscriber Loop includes such technologies as ADSL 1GE 1 Gigabit Ethernet 10GE 10 Gigabit Ethernet Date 30 11 2009 Page 130 The following diagrams illustrate the level of configuration detail contained within the revised model v3 1 for each of the major equipment categories DSLAM MSANs Maximum customer connections Max fill factor final line card 60 Max fill factor overall DSLAM 95 Chassis Type 1 Slots for cards 4 Ports per card 48 Maximum ports fully configured 192 Maximum ports allowance for growth 182 Percent of costs traffic related 20 Chassis Type 2 Slots for cards 16 Ports per card 48 Maximum ports fully configured 768 Maximum ports allowance for growth 729 Percent of costs traffic related 20 Usage Drivers Allowance for Spares Chassis card related Subs 1 0 Chassis traffic related Mpps 1 0 POTS card Subs 5 0 ADSL card Subs 5 0 VDSL card Subs 5 0 Date 30 11 2009 Page 131 Layer 2 Aggregation
5. The Scorched node approach implies that gt At least one switch or router is placed at each exchange gt All transmission and switching equipment is co located on the site No sites in addition to the existing sites are required gt No sites are removed all existing sites remain in the network including technical houses and gt The size of the building required to contain the equipment is determined in the model The trenches and cables connecting the sites are routed optimally It is assumed that these trenches and facilities can be shared with other utilities and also with the access network 9 1 Overview of the Core Model 9 1 1 Network structure As has already been mentioned NITA is concerned that the Hybrid Model should continue to reflect an efficient operator utilising an optimally configured network based on latest proven technology Therefore NITA believes that from version 2 5 onwards the underlying technology on which the model is based should no longer be based on circuit switching but instead packet switching and in particular an all IP core This core network would have all services running across it from traditional telephony to more advanced triple play services utilising for example multicast and multichannel functionality The basic concept of an all IP core network is relatively straightforward and indeed was addressed during the 2006 version 2 4 revision of the NITA model The hierarchical n
6. between the centroid of the PDP zone and the PDP cabinet the location strategy of the PDP cabinet plays an important role here given that centroid of the zone and ASM are assumed to connected through a straight line From PDP to SDP This is worked out through trigonometric formulas if the preferred approach is consistent with the other assumptions in the model and the usage of the elliptic function if the fall back approach assuming circular SDP areas are used instead A direct examination of the formulas is the best way of understanding this part of the model which is quite complicated From SDP to EFSD This distance in the preferred approach is worked out on the basis of street length and width whereas in the fallback approach on the elliptic function calculations the fallback approach allows for this methodology to be applied because the SDP area is assumed to be circular From EFSD to NTP This is worked out as the summation of two distances from street duct to building from building to NTP The inputs used come directly from the data on buildings and street sizes for each of the 20 sample zones An input which exerts an influence on the location of PDPs and SDP and the length of the different parts of the subscriber line is the location strategy This input is applied to assess the relative location of the PDPs and the SDPs between the ASM and a randomly selected subscriber in the area The value of the parameter is between
7. input table worksheet I Route _Table A product can have up to three separate routes defined with percentag likelihoods attached to each one The weighted averages of these separate routes are collated together in a table at the bottom of that worksheet The routing factors are combined with the product demand in worksheet C_Route_Table and it is this worksheet that aggregates the demand for each network element This aggregate network element demand is then overlaid onto the network structure in worksheets I Trenching Transit I Trenching Local and I Node Equipment in order to dimension the equipment at each layer of the network hierarchy taking into account the aggregate demand on that network element together with the number of instances of the element in the network and the relevant dimensioning rules defined in worksheet I Technical When the results for the individual services are summed in the Consolidation Model it is visible how much the different parts of the network are utilised to deliver fixed network services 9 3 2 1 Busy Hour conversion The values for Erlang busy hour are an important element of network dimensioning The model uses the conversion factor BHE yearly number of minutes 52 busy days per week busy hours per day 60 to convert the yearly number of minutes to Erlang in busy hour The yearly traffic load is first split by a factor of 52 for the number of weeks per year In the next step traffic is split by t
8. where appropriate Date 30 11 2009 Page 49 6 1 Introduction The 2009 release of the Hybrid Model contains a lot of changes most fundamental of which has been the decision to base the core network on an all IP network structure This in turn led to a need for the demand associated with additional products to flow through the model to the output sheets in the consolidation model Since this greatly increased the number of products addressed NITA decided to have two output sheets The first of these O_Output would contain data related to all products modelled whereas the second O Results would only contain data related to regulated products and services The changes made are summarised in the following sections 6 2 Main LRAIC service costs O_Output Core Services now covers all voice related products including those connecting to traditional customers and those connecting to customers utilising TDC s VoIP service Additional products have been added to Colocation Costs These encompass additional power variants Access Services now covers all access related products These also show prices broken down by geotype Bitstream Access has been extended to include BSA National Access and also to show prices broken down by geotype Additional sections have been added to cater for Bitstream Transport IPTV and VoD and costs allocated to residual non PSTN products O_Results Core Services now only includes Local Intercon
9. 2 5 Pre selection 117 Date 30 11 2009 10 2 6 Implementation of prefix for pre selection117 10 2 7 Interconnection 117 10 2 8 Supervised access with notice 118 10 3 How to use the Co location Model 119 10 3 1 Model scope 119 10 3 2 Model structure 119 10 3 3 Start up 120 10 3 4 Understanding the model 120 11 The Consolidation Model 121 11 1 Overview of the Consolidation Model 121 11 2 Operating costs The Functional Area approach 121 11 2 1 Functional Area inputs 122 11 2 2 Staff costs 122 11 2 3 Calculation of pay costs 122 11 2 4 Calculation of total costs 123 11 2 5 Overheads 123 11 3 Calculation of service costs 123 11 3 1 Annualised costs by Network Element 124 11 3 2 Basic product costs 124 11 3 3 Final product costs including mark ups 124 11 4 How to use the Consolidation Model 124 11 4 1 Model scope 124 11 4 2 Model structure 125 11 4 3 Start up 127 11 4 4 Understanding the model 127 Page 5 12 Appendix A Abbreviations 128 13 Appendix B Core Model Inputs 131 14 Appendix C Worksheet Audit 135 a args smear satiate se 14 1 Description of worksheets in the Access 22 20 n errr rrr Model 135 14 1 1 Input sheets 135 14 1 2 Calculation sheets 137 14 1 3 Output sheets 138 14 2 Description of worksheets in the Core Model 138 14 2 1 Utility sheets 138 14 2 2 Input sheets 138 14 2 3 Calculation sheets 140 14 2 4 Output sheets 141 14 3 Description of worksheets in the Co location Model 141 14
10. 2010 and the years following meaning that the prices for 2010 are based on data for the year 2009 The costs are annualised in order to ensure that they are recovered over the economic lifetime of the asset In previous versions the Hybrid Model used a tilted annuity which calculates the sum of depreciation and cost of capital In version 2 5 of the Hybrid Model users are able to select from a range of annualisation methods however the approach selected by NITA and used as the default setting is a still tilted annuities 4 3 Calculation of economic depreciation The annualisation method used in the Hybrid Model should be consistent with two principles gt Each asset cost is fully recovered for over its own lifetime gt An existing network operator is able to compete on an equal basis with a hypothetical new operator entering the market in a later year on equivalent terms In addition annualisation methods should ideally reflect all of the following gt An underlying price trend in the purchase price of the asset gt An underlying output trend in the number of units of output produced per year gt An underlying change in the reliability of the asset and hence need for opex to maintain the asset Depending on different economic conditions both tilted annuity and economic depreciation can meet these requirements and therefore the Hybrid Model can make use of both methods The conditions identified by NITA as influencing
11. 30 11 2009 Page 122 11 2 4 Calculation of total costs This section combines the staffing costs from the previous section with the non pay costs from the third section to calculate the total costs for each functional area It also in table 4 3 of the worksheet groups together the network related operating costs into one of a number of broad areas including for completeness a number of other operating costs these being gt Supplier annual support costs gt Annualised accommodation costs gt Power costs gt Air conditioning costs Table 4 3 of the worksheet thus provides a useful summary of the aggregate operating costs of the overall network 11 2 5 Overheads This section calculates a number of mark ups used in worksheet C_Services The mark ups comprise three types gt General mark ups for common business costs gt Specific interconnect and commercial costs gt NP IN costs The general mark ups are calculated as a blend of an overall mark up for common business costs with specific mark ups for working capital The former is based on the total non network costs from Table 4 4 in the worksheet divided by the total cost base calculated by the model mainly from worksheet C_Costs but including some adjustments from worksheet C_Services The working capital mark ups are specific to access and core and are based on a number of inputs contained within this sub section The mark ups for specific interconnect and commercial c
12. 6 Access fibre 5 5 Cabinets distribution points 1 1 NTPs 0 0 Line cards 0 6 Circuit boards jointing 1 1 DSLAMs MSANs 0 0 Layer 2 switches 0 7 5 Layer 3 routers 0 Peering routers 0 Media gateways 0 IMS controllers servers 0 IN platform 8 8 MDF 0 0 International switch 6 6 DWDM transmission 0 Power supply unit 0 0 Backup power 4 4 Air conditioning unit 2 2 Security system 0 0 Security guard 0 0 Site preparation 0 0 Site maintenance 0 0 Core duct 3 3 Core trench 3 3 Core cable 5 5 Submarine cable 8 8 Technical house rack space 0 0 POI 0 0 Buildings 0 0 Land 4 4 4 6 1 Trench and duct in access and primary networks Statistics Denmark regularly publishes an index concerning the regulation of tenders for earth and asphalt work The view of NITA is that this index provides a good indicator for the price of digging trenches Date 30 11 2009 Page 41 The official statistics for 2001 2005 show a positive price trend of between 3 1 and 3 6 for earth and asphalt work Based on this NITA used a price trend of 3 in the 2005 LRAIC hybrid model for trench in the access and primary network This price trend was also used in the original version of the LRAIC Model and is supported by international benchmarks For the 2009 version updated statistics were obtained together with actual data from the LRAIC Working Group relating to digging costs which ha
13. C_Services be updated as a result then it is very likely that these inputs will also need to be re visited if only to ensure that they remain relevant for the network elements for which they were intended 11 3 2 Basic product costs This section comprises seven parts The first of these allocates the core network element costs between the various products addressed in the core model Note that for any particular product the costs will include costs related to Gbps Mpps calls minutes and subscribers The remainder of this section contains six parts related to the access network elements The first of these calculates the total usage of each network element by each product This is calculated for each geotype individually and for the country as a whole The other five serve to allocate the network element costs on the basis of this usage with one part addressing the country as a whole and the following four parts addressing each of the four geotypes in turn 11 3 3 Final product costs including mark ups This section collates the results of the previous sections applies relevant mark ups and calculates per unit costs as appropriate It comprises three parts dealing with core products access products and co location products in turn 11 4 How to use the Consolidation Model 11 4 1 Model scope The purpose of the consolidation model is to carry out the final steps in the costing process and allocate costs to the final services The con
14. Consolidation model is provided in Appendix C of this document Date 30 11 2009 Page 127 ADM ADSL ASM ATM BH BHCA BSA Codec CSCF DSLAM DWDM eBSA EU FTE FTTB FTTH GBIC Gbps GW IMS IP IPTV ISDN Kbps LLU Date 30 11 2009 Add Drop Multiplexer Asymmetrical Digital Subscriber Line Remote Concentrator site Asynchronous Transfer Mode Busy Hour Busy Hour Call Attempt Bitstream Access Coder Decoder Call Session Control Function Digital Subscriber Line Access Multiplexer Dense Wavelength Division Multiplexing Ethernet Bitstream Access European Union Full Time Equivalent employee Fibre to the Building Fibre to the Home Gigabit interface card Gigabits per second Gateway Integrated Multimedia Subsystem Internet Protocol IP Television Integrated Services Digital Network Kilo bits per second Local Loop Unbundling Page 128 LRAIC Long Run Average Incremental Cost L2 Layer 2 as in 7 layer ISO model L3 Layer 3 as in 7 layer ISO model MDF Main Distribution Frame Min Minute Mpps Million packets per second MRP Model Reference Paper MSAN Multi Service Access Node NGA Next Generation Access NGN Next Generation Network NPV Net Present Value NTP Network Termination Point OAO Other Authorised Operator Opex Operating costs PDP Primary Distribution Point POI Point of Interconnect PON Passive Optical Network POTS Plain Old Telephony Service PPS Packets per second
15. If we assume that a second operator enters the market with identical assets and that these assets are annualised by tilted annuity or economic depreciation so that their depreciation charges equal those of an existing operator then the second entrant would still be able to undercut the incumbent simply because it would benefit from newer assets and lower associated operating costs This violates the modelling principle that both entrants should have the same efficient cost base assuming identical assets To put both operators on an equal footing the model makes the following assumption Just as an operator would choose to depreciate its assets more heavily in the years before Operator 2 enters the market so that it has a lower cost base when competition arrives the principle underlying tilted annuities it might choose to bring forward some of the anticipated opex associated with ageing assets to guard against the threat of being undercut by a new entrant If Operator 2 enters the market on the same basis and smoothes its own opex to be prepared for future competition then its cost base on entry will be the same as that of Operator 1 This is modelled as follows gt Assume that opex increases by a given percentage every year due to the underlying asset getting progressively less reliable gt Find the NPV of the opex cashflows over the lifetime of the asset gt Turn that NPV into a constant ie non tilted annuity This results in
16. Model 75 8 6 Next Generation Access 75 8 7 How to use the Access Model 76 8 7 1 Model scope 76 8 7 2 Model structure 77 8 7 3 Start up 77 8 7 4 Understanding the model 78 9 The Core Model 79 9 1 Overview of the Core Model 79 9 1 1 Network structure 79 9 1 2 Scope of the core model 80 9 2 IP network design 80 9 3 Network dimensioning 86 9 3 1 Basic assumptions 86 9 3 2 Method for network dimensioning 89 9 4 Key Network Elements 90 9 5 Core trenching 91 9 6 Modelling the core network 93 9 6 1 Voice telephony overlay 93 Date 30 11 2009 Page 4 9 6 2 DWDM 94 9 6 3 Equipment configuration 95 PEE rine nates tabbed SON A 9 7 Network demand parameters 96 9 8 Detailed description of cost calculations 97 9 8 1 Overview 97 9 8 2 Cost input 98 9 8 3 Cost allocation methods 99 9 8 4 Calculation of service costs 100 9 9 How to use the Core Model 101 9 9 1 Model scope 101 9 9 2 Model structure 101 9 9 3 Start up 102 9 9 4 Understanding the model 102 10 The Co location Model incl Other Services 104 10 1 Co location services 104 10 1 1 Definition 104 10 1 2 Modelling Co location 105 10 1 3 Cost categories 107 10 1 4 Power supply 108 10 2 Other services in the Hybrid Model 110 10 2 1 General assumptions about hourly rate and overhead cost 112 10 2 2 Handling double counting of order processing and transport 112 10 2 3 Raw copper and shared raw copper 113 10 2 4 BSA installation 116 10
17. Scope of the core model The core model is a bottom up model of the equipment and systems required to carry the services defined in the MRP with the required level of service quality It therefore calculates the cost of both wholesale interconnect and retail PSTN and IP Network products The core network model is defined to include all systems and equipment contained in scorched nodes including links between the nodes It does not include links from the node to the customer these links are in the access model Whilst the core network increment does not include line cards and the MDF these are still included within the core model for ease of modelling The costs associated with these items are then allocated to the relevant access products in the consolidation model The core model includes additional costs required to give the extra capacity needed to support other services than those defined in the MRP These other services include IPTV leased lines and non broadband Bitstream datacom services The costs of servicing and supply of dedicated equipment for these increments are not included However the capacity effect on the PSTN IP Network increment is taken into account Due to the effect of cost volume relationships this increased capacity reduces the average cost per unit of products and services using that capacity 9 2 IP network design The basic design of an IP network is a hierarchical one as depicted in Figure 9 1 below Dat
18. Switches Configuration Rules Maximum number of sites per Layer 2 ring Ss 10GE ports per chassis for L2 ring 2 10GE uplinks per chassis at L3 site a Maximum number of L3 routers fed by Bitstream uplirn 2 Maximum utilisation of a L2 ring _ s 7 Gbps Maximum number of GBIC modules per chassis sg Chassis selection Maximum xDSL customers for lower specType1cha 1000 Processor selection Maximum ring traffic for lower spec Processor Type1 2 Gbps Maximum xDSL customers before Redundant Proces Processor Type 1 Number of 10GE ports Number of GBIC ports Processor Type 2 Number of 10GE ports 1500 2 eee 2l Number of GBIC ports o _ l 2 _ 48 10GE Module Number of 10GE ports GBIC Module Type 1 Number of GBIC ports Maximum number of DSLAMs per module GBIC Module Type 2 Number of GBIC ports Usage Drivers Allowance for Spares Chassis Type 1 Mpps 1 0 Chassis Type 2 Mpps 1 0 Processor Type 1 Mpps 1 0 Processor Type 2 Mpps 1 0 10GE Module Mpps 5 0 GBIC Module Type 1 Mpps 5 0 GBIC Module Type 2 Mpps 5 0 10GE Adaptor Gbps 5 0 GBIC Adaptor Gbps 5 0 Date 30 11 2009 Page 132 Layer 3 Edge Routers Configuration Rules Minimum number of chassis per site Maximum number of modules per chassis Maximum number of 10GE per module Maximum number of 1GE per module Allow 1GE ports for L2 links Usage Drivers Allowance for Spares Chassis Mpps 1 0 10GE Modules Mpps 5 0 10GE ports Gbps 5 0 1
19. Tag L 4 Bytes Ethernet Preamble 8 Bytes Ethernet Header ss 14 Bytes Ethernet CRC _ 4 Bytes Ethernet Frame Gap _ ss 12 Bytes RTP Header Compression No SeeRFC2508 cRTP potential saving 36 Bytes Voice activity detection VAD No VAD bandwidth saving C n Uplift to account for data flows related to call set up cpe registration polling etc Packet rate for one voice channel pps 51 00 IP bandwidth for one voice channel Kbps 98 74 Upstream bits per second as percentage of downstream Upstream packets per second as percentage of downstream The second issue relates to the average size in terms of bytes of IP packets for one service compared to another The reasoning behind this is that much of the IP related equipment will process for example route the data on the basis of complete packets rather than individual bytes The model therefore needs to incorporate a number of bits per second to packets per second conversion factors such that the costs of relevant equipment can be allocated correctly The conversion factors included are illustrated below with the exception of voice telephony which was shown in Figure 9 11 Date 30 11 2009 Page 96 Figure 9 12 IP Packets Conversion Factors Broadband parameters Average size of broadband packets Bytes 750 Million pps per Gbps conversion 0 17 Header total 62 Bit Rate uplift due to headers 8 3 Upstream bits per second as percentage of downstream 25 Upstream packe
20. a sense check or some calculations have given results that do not appear to be correct The error message will indicate the nature of the problem but it is left to the user to resolve the problem gt Cells that link to external data Tan coloured areas indicate linked data from another workbook gt Confidential input data Bright blue indicates a confidential input The user may enter a value or occasionally enter a hard coded formula to create the value required this is not recommended but it can be useful Values in these cells have been camouflaged for reasons of confidentiality gt Single cell array calculations These are often accompanied by a warning message The inexperienced user should particularly avoid editing these cells as single cell array calculations can get corrupted by being edited even if the user does not change the formula These array calculations are used to perform larger calculations whilst keeping the size of the model manageable Refer to the Excel on line help for more information on array calculations Excel Comments and some floating text boxes have been used to expand on the user notes Some are in colour simply to highlight the information more clearly Comments fields also supply information about specific data or formulae used The comments and explanations within the model are a vital part of the model since they aid the user in understanding the model Hence the model can only be fully understo
21. after discounting recover the asset s purchase price and financing costs 4 1 5 Economic depreciation The main drawback of the tilted annuity method is that it ignores output trends This is addressed by economic depreciation the method which most closely reflects economic reality and is widely considered the theoretically correct annualisation method to use in LRAIC models Economic depreciation is defined as the period by period change in the market value of an asset The market value of an asset is equal to the present value of the net cashflows that the asset is expected to generate over the Date 30 11 2009 Page 33 remainder of its useful life As net cashflows vary with output assets are depreciated at a rate consistent with use resulting in a true depreciation profile As with tilted annuities economic depreciation ensures that two entrants have the same depreciation charges regardless of when their assets were acquired 4 2 Annualisation methods used in the Hybrid Model According to the legislation NITA will set the prices using the LRAIC Hybrid Model for one year at a time During the first year in which the LRAIC prices were determined 2003 the Hybrid Model produced results based on information relevant for the year 2002 Prices for the following years were determined on updated versions of the model reflecting any changes in equipment prices and volumes This principle will also be used for the price estimation for
22. cost data from the core access and co location models Each cost category is assigned a network element that relates to the equipment costs installation costs and operating costs provided by the three models Further price Date 30 11 2009 Page 143 trends projected price change asset lives scrap values output trends and other parameters are imported to facilitate the annualisation of costs There are a number of input changes that may be performed in this sheet light blue cells cf the style overview They include gt Allocation All input costs must be defined to have an allocation this specifies how the cost category taken from any of the three models is to be allocated Thus the cost category DWDM Base Units Core Core maps to the network element TX path L3 Core Core and Acc Copper 2 pair maps the cost to the element Acc Cable Copper SDP NTP gt Expensed vs annualised All cost inputs must have a specification of whether the cost should be expensed or annualised gt Annualisation option The model allows the user to select from a menu of four different annualisation methods straight line standard annuity tilted annuity and output adjusted tilted annuity economic depreciation To manually select a method for an individual asset the user can use the override column Column AP otherwise the model applies the calculated default method for economic depreciation It is also pos
23. costs are also common costs Secondly there are certain fixed costs associated with host and remote subscriber stages and DSLAMs such as installation costs Again any knock on costs associated with these common fixed costs are also common fixed costs The treatment of common costs in the Hybrid Model is described in Chapter 3 Fixed costs are defined here as costs that do not change with the level of output Date 30 11 2009 Page 11 Box 1 The Relationship Between Cost Concepts Directly attributable Directly attributable costs in the access costs in the core network e g MDF network e g tandem exchange Shared costs in the Common costs e g Shared costs in the core access network e g trench shared by access network e g trench in core trench in access network and core network ADMs etc AN A Access Network Costs Core Network Costs Directly attributable costs are those costs that are incurred as a direct result of the provision of a particular service in a particular increment These costs fall into two types First the costs of some inputs vary with the level of output so that even if the output of more than one service requires this input the extent to which a single service causes the costs can be calculated Second there are assets and operating costs which are fixed with respect to the level of output but which are service specific Shared costs are the costs of those inputs necessary to produce
24. costs such as the line card within the active network equipment for example the concentrator and or DSLAM This is consistent with the first view since for example line card requirements are driven by the number of subscribers or more accurately by the subscriber requirements for lines It is also consistent with the second view since the line card is essential to sending and receiving traffic Assets within the access network typically include gt The dropwire to the subscriber gt Trench including any duct between the concentrator and the subscriber s premises gt Street cabinets gt Copper cable and optical fibre in this part of the network gt Other assets such as manholes poles and overhead cables if used etc gt Network Termination Points NTP gt Subscriber related costs such as line cards in the active network equipment such as concentrators DSLAMs Date 30 11 2009 Page 13 2 1 3 Other increments In practice there will be several other increments in addition to the core and access increment Examples include a retail increment for the access and core networks an increment for premium rate services an increment for the mobile network and an increment for other services In the LRAIC Hybrid Model these other increments are not modelled explicitly The model identifies the costs that are common between other increments and the core and access network increments 2 2 Other main assumptions T
25. currently has a significant number of customers using its existing VoIP service This in turn means that TDC already has in place a network structure capable of handling such traffic and of interconnecting such traffic with a legacy TDM based network The conceptual design adopted for the revised model follows the same broad principles as that currently used by TDC Date 30 11 2009 Page 93 IMS Session Signalling IMS User Plane Data It follows the broad IMS structural guidelines as illustrated in Figure 9 9 above A number of CSCF Call Session Control Function softswitch servers are located across the network One such server facility has been assumed at each of the 13 Distribution router sites though clearly other configurations would be equally viable All voice telephony traffic is fed through a Session Border Controller for reasons of seeking to maximise the overall integrity of the service A Session Border Controller is located again at each of the Distribution node sites Interconnection with TDM based networks is via IP TDM Media Gateways themselves controlled by Media Gateway Controllers These have been assumed to be located at each of the Distribution node sites and also at each of the Edge Router node sites 9 6 2 DWDM Dense Wavelength Division Multiplexing DWDM equipment is currently used within TDC s network for one of three principal reasons gt To address the need to achieve point to point connect
26. equal annual opex charges over the lifetime of the asset These theoretical cashflows can be then be modelled in the same way as capital expenditure and adjusted for output as appropriate Date 30 11 2009 Page 37 4 4 The Cost of Capital The cost of capital in the LRAIC Hybrid Model is based on the weighted average cost of capital WACC of an efficient Danish SMP provider A telecom company is financed both by equity and loan capital Consequently the total cost of capital is the average cost of utilising the two types of capital The weighted average cost of capital comprises the company s combined cost of equity and debt WACC before tax is calculated as E D WACCo3eFo s R e R e where BEFORE TAXE F p FD gt Eis equity gt D is loan capital or debt gt Reis the return on equity gt Rp is the return on loan capital The cost of capital used in version 2 5 of revised LRAIC Hybrid Model for both the core and access networks is 8 0 4 5 Asset lives In the LRAIC Hybrid Model the asset lives should reflect the economic asset life This is the period in which the asset s earnings exceed its costs It is important to note that the physical life of the asset generally will be longer than the economic life The column on the right hand side of Table 4 2 below depicts the asset lives used in the LRAIC Hybrid Model As the estimates of the economic asset lifetimes are not integers values have generally be
27. future cashflows generated by 5 the asset m Q Another way of looking at this is that if o the market price limits what can be charged for a product then the company would tend to keep a static unit price over the years which would equate to lower depreciation in the early years No advantage gained by later entrant as Makes good business sense to they still have to buy the same quantity accelerate depreciation as later entrant Of the asset over time since all costs are can price lower since does not need to 9 fixed i e do not depend on output buy as much asset i e it is not a trends constrained by previous decisions oc zg D If there are market and or regulatory Makes good business sense to constraints on the product market price accelerate depreciation as later entrant o and output is falling then the can price lower since does not need to g depreciation charge should be reduced buy as much asset i e it is not Z over time to reflect the lower value of constrained by previous decisions a future cashflows generated by the F asset Another way of looking at this is that if oO the market price limits what can be charged then the company would tend Economic depreciation may be modelled explicitly by forecasting capex opex asset prices and output year on year over the full economic lives of network assets which for some assets eg trench copper may be 30 years or longer Altho
28. in Denmark to refer to the products purchased by OAOs that they can then use in conjunction with their own network infrastructure to provide retail xDSL services in competition with TDC Various forms of BSA are possible depending on the level of interconnection in TDC s network TDC s network originally used ATM based DSLAMs with OAOs connecting either at a national level utilising TDC s ATM network for backhaul from the DSLAMs to the single point of interconnect or at a regional level at the first ATM node As of the end of 2005 less than 1 per cent of BSA based end users were routed to the OAOs at the ATM node In 2006 TDC embarked on a programme of deploying DSLAMs deeper into the access network than at the ASM level and announced plans to move around 1 500 DSLAMs into Street Cabinets within the next few years although the Date 30 11 2009 Page 46 actual number in place at the end of 2008 was much less than this The DSLAMs used for this programme are modern Ethernet based ones and TDC also expects over time that the existing ATM based ones will also be upgraded or replaced by this current technology although the precise timeframe for this is known Figure 5 2 Schematic illustration of Ethernet based xDSL network IP MPLS core router IP MPLS core e ISP Router DSLAM L2 switch ISP Router ISP Router ISP Router The network structure relevant to the Ethernet based DSLAMs consists of a number of l
29. in other categories such as PSTN or ISDN line rentals The purpose is to estimate demand for all access lines avoiding double counting 2 3 2 Leased lines Users of leased lines may be classified in the following three groups gt Retail subscribers who usually require leased lines to provide a permanent connection between subscriber premises gt Other operators who usually require leased lines to provide a permanent connection between networks gt The network operator who requires leased lines for a variety of reasons SMP operators may carry some other services such as data services over leased lines 2 3 3 Other services Other services using the core network will increase in importance over time Examples include IPTV packet switched Virtual Private Networks VPN and video on demand The costs of servicing and supply of dedicated equipment for these services are not modelled However the capacity effect on the PSTN IP Core Network increment is taken into account Due to the effect of cost volume relationships this increased capacity reduces the average cost per unit of products and services using that capacity Date 30 11 2009 Page 16 This chapter describes the assumptions made by NITA regarding unit costs direct and indirect and operating costs This is followed by a chapter discussing the assumptions and the approach used to annualise these costs 3 1 Defining costs There are two methods of defining costs w
30. in the Layer 2 Ethernet Switch gt Ring facing line card in the Layer 2 Ethernet Switch gt Layer 2 ring transmission link gt Layer 3 Edge Router facing line card in the Layer 2 Ethernet Switch gt Layer 2 Ethernet Switch facing line card in the Layer 3 Edge Router gt Processor in the Layer 3 Edge Router gt Ring facing line card in the Layer 3 Edge Router gt Layer 3 Edge ring transmission link gt Layer 3 Distribution Router facing line card in the Layer 3 Edge Router gt Layer 3 Edge Router facing line card in the Layer 3 Distribution Router gt Processor in the Layer 3 Distribution Router gt Session Border Controller at the Layer 3 Distribution Router site gt IMS Core at the Layer 3 Distribution Router site onmly used for call set up gt A second instance of the Layer 3 Edge Router facing line card in the Layer 3 Distribution Router gt A second instance of the Layer 3 Distribution Router facing line card in the Layer 3 Edge Router gt A second instance of the Layer 3 Edge ring transmission link gt A second instance of the Ring facing line card in the Layer 3 Edge Router Date 30 11 2009 Page 87 gt A second instance of the Processor in the Layer 3 Edge Router gt A second instance of the Layer 2 Ethernet Switch facing line card in the Layer 3 Edge Router gt A second instance of the Layer 3 Edge Router facing line card in the Layer 2 Ethernet Switch gt A second instance of the Layer 2 ring
31. minutes using the element However costs have been annualised in order to allocate costs to services This is done in the consolidation model where annualised GRCs and installation costs are added to operating costs where the FA correction has been applied in order to derive the annual cost Cost of capital asset lifetime and the price development are combined in this calculation As a final step the costs of services are uplifted with one or more mark ups for overhead and costs related to working capital Date 30 11 2009 Page 100 9 9 How to use the Core Model 9 9 1 Model scope The purpose of the core model is to calculate the resources required to build a switching and transmission network for an operator of the size and scope of TDC As discussed previously this is based on an all IP network structure rather than the existing mix of platforms in the TDC network The main components of this are gt Cost of IP switching and routing equipment gt Cost of sites gt Cost of trenching gt Cost of transmission infrastructure gt Certain types of access network costs In relation to the last of these components the core model calculates the costs of that part of the access network located at exchange sites primarily line cards and MDF equipment The core model incorporates a large number of nodes physical sites interconnected by transmission paths The majority of these nodes are considered to form part of the cor
32. national weighted output NITA has taken note of the LRAIC Working Group s desire for regulated access to backhaul bandwidth and in particular for its inclusion within the next revision of the Hybrid model NITA has initiated a formal analysis of the markets for both unbundled access and broadband connections in accordance with the EU Commission s new recommendations on Markets 4 and 5 The revised model therefore has been adapted to help anticipate potential regulated access backhaul products and add the necessary functionality to allow for the calculation of cost based prices should that prove necessary 8 2 3 Key network elements The main purpose of the access model is to calculate the cost of raw copper access defined as access to the copper line stretching from the Main Distribution Frame MDF to the customer premises The access model itself does not calculate access network costs essentially customer facing line cards at the DSLAM MSAN site or the MDFs at those sites these are calculated within the core model The main components of the access model are gt Trench Duct and Poles used by the access network gt Copper cables gt Fibre cables gt Passive street cabinets and distribution points gt Final drop miniduct duct to subscriber building from street associated digging works and Network Termination Points NTPs gt Active street cabinets but not the equipment inside them gt Backhaul trench duc
33. number of line cards and therefore their total cost will depend on the number of subscribers In contrast an indirect network cost is one where the volume of inputs and hence cost depends on choices made concerning other direct inputs An example is racks since the number and size of racks necessary will depend on the choices made concerning ports and line cards Interconnection specific costs comprise costs which may be associated with supplying interconnection services at the wholesale level These are typically subscriber related costs such as billing and billing system costs Overheads cover those costs that are not necessary to run a network but must nonetheless be incurred in order for the network operator to function Date 30 11 2009 Page 17 The approach used to estimate these types of costs is presented below Table 3 1 shows how the different costs have been treated in the Hybrid Model Table 3 1 Definition of Costs Used in the Hybrid Model Cost Category Examples of Costs Included Direct costs Trench Line cards Copper cable Ports Fibre Indirect capital costs Accommodation Power Security Air conditioning Operating costs direct Installation operation and maintenance of the network FA costs Operating costs indirect Travel and management costs Overheads Executive Business planning and development Accounting finance and audit External relations Hu
34. of room fit out with a cost of materials of DKK 64 per rack and installation costs of 1 3 hour for a technician per rack space On the basis of the above inputs NITA has estimated the costs for room fit out excl lighting and fire alarms at DKK 2 500 and for cable trenches at DKK 7 000 per rack space 10 1 4 Power supply NITA has individually modelled the different costs related to installation and the annual consumption for four power supply services Date 30 11 2009 Page 108 For each of these the following parameters are defined gt Materials gt Number of hours for technicians physical installation gt Annual consumption in kWh A separate category calculates the cost of order processing exchange work and transport time to exchange For all power supply services the cost per kWh is applied Given the cost of materials and hourly salary costs for administrative and technical staff the total installation cost for each power supply service can be calculated In addition given the cost per kWh and an assumption of an average power consumption it is possible to calculate the ongoing annual cost The results for power supply in the hybrid model are split between a one off cost and an annual cost for the power supplies Power 48 VDC 16 A Power 48 VDC 30 A Power 48 VDC 60 A Power 230 VAC 10 A and Power 230 VAC service socket In response to an industry consultation in 2008 the service shift to another power
35. of the access network that are at or above the existing MDF site are dimensioned and costed along with operations and maintenance costs in the core model and results costs and resources used are passed to the consolidation model The main components of the access model are gt Trench and duct used by the access network gt Copper cables and distribution points including Active Street Cabinets gt Final drop miniduct associated digging works and Network Termination Points NTPs Date 30 11 2009 Page 76 gt Fibre in the access network gt Backhaul from Active Street Cabinets After calculating the resources needed to build this part of the network the model uses the results to calculate the capital cost of those resources and the annual operating cost of such a network 8 7 2 Model structure For ease of navigation the access model includes a map of all major information flows between them This is reproduced in the diagram below Figure 8 5 Access Model Navigation Map Input Calculations Output _Access Routing Table _Share Routes C_Share Routes _GIS Trench and C_Trench and C_Trench by O_To Duct Duct Zone Consolidation _Network _GIS Zones C_Cables and C Resources Demand _Cost of Nodes O_Key Statistics erence _ASM PDP _ Opex Costs til alle ark stikpr ve p 10 centralzoner NB For at opdatere resultateme for Cable and Nodes skal der udf res en makro for at
36. opdatere resultaterne for en The user can click on any of the buttons on the map to jump straight to the relevant worksheet Each subsidiary worksheet has a Model Overview button which the user can click on to return to the navigation map Most sheets also have a drop down menu below the Model Overview button to take the user to different tables contained in each worksheet Note that it is still possible to navigate the model using the sheet tabs The arrows show the major flow of information between the different sheets to give an illustration of the logic of the access model 8 7 3 Start up There are two ways to open the Access model The first method which is recommended is to first open the Consolidation model and then select the Open Linked Sheets option which will open all the individual worksheets of the Hybrid Model This ensures that linked information is updated automatically and links are preserved when the model is saved under a different filename The user should also be aware that the Excel auditing tools will trace precedent and dependent cells only for workbooks that are currently open Date 30 11 2009 Page 77 Alternatively the user may open the Access model directly from the file directory Enable macros when asked unless the file is from an uncertain untrusted source There may be a message asking whether to update links Such updates are performed automatically on recalculation if other parts of th
37. operating costs Local points of interconnect include the estimated cost of the service interconnection bundles that is included as an individual element in TDC s standard offers An interconnection bundle is defined as group of lines or 30 groups connecting the same pair of exchanges with similar technical environment such as signalling and direction However because TDC previously informed NITA that normally only one interconnection bundle per point of interconnect is installed NITA has not found it reasonable to investigate this issue further 10 2 7 1 Local Point of Interconnect incl interconnection bundles Installation The cost of installing a local point of interconnect covers the following activities gt Processing of order includes receiving forms for traffic orders and orders for interconnect and transmission putting orders in Columbus and processing the confirmation gt Changes in the network includes drawing up the signalling parameters operation and maintenance e g surveillance functionality and traffic data including programming This is assumed as tasks for technicians Date 30 11 2009 Page 117 Annual cost The annual cost of a local point of interconnection covers the following activities gt Administration including both administrative personnel e g billing secretaries and academic personnel case processing gt Network planning contains operating and maintenance tasks such as
38. physical re arrangement is a service that is carried out in connection with the installation the transport is already covered by this service Transportation is widely covered by the installation Additionally it is understood that the technicians are based locally which also reduces the need for transport Therefore a correction for double counting of time for transport the share covered by physical rearrangement is done in relation to the services installation of raw copper shared raw copper and other conditions that reduce the need for transportation 10 2 3 6 Change of trunk number The cost of change of trunk number is modelled from the following categories 1 Processing of order consists of cost of ordering before and after the physical coupling in the exchange e g reception putting the order into the system and re registration of trench 2 Physical coupling in exchange covers actually carrying out the order on the exchange including connection the cross thread coupling to trunk and securing that there is connection to the operator s co location 3 Transportation includes the transport to and from the exchange Time for transport includes transport to the un manned exchanges It can be reduced when the technician can carry out orders at one exchange at the same time 10 2 3 7 Visit by technician unproductive visit by technician When a visit by a technician is billed by the hour and is carried out within nor
39. sites was separately assessed and measured with Autoroute The resultant trench network is illustrated in Figure 9 8 below The blue routes are those necessary to interconnect the higher level sites Distribution router and above and the red routes the additional routes to include the Edge router sites The purple routes are those that need to travel underwater Figure 9 8 Core Trench Mapping in Autoroute Undersea cable routes were included where necessary to interconnect islands though bridges were used in preference where available For the lower level nodes ie connections between the DSLAM MSAN sites and the Edge routers each Edge router area was separately assessed and measured with Autoroute though the individual route segments were not separately recorded within Date 30 11 2009 Page 92 the model due to the large number involved Instead only aggregate distances were included Once the core trench routes had been established the necessary logical routes could be overlaid onto them This was done for each level in the network hierarchy following the broad principles adopted by TDC in its own IP network design This consists of gt 4 core node sites each containing two routers interconnected in a square arrangement gt 13 distribution node sites 4 shared with the core node sites each containing two routers with each site having direct connections to two core node sites but not having connections to eac
40. such as support contracts and equipment suppliers The bottom up guidelines indicate that a percentage increase in the debtor days may be used instead of a figure for the amount of cash required by a prudent operator Cash could also be defined as a percentage increase in sales revenue in the equation 2 above and these are seen as being equivalent The working capital given by company accounts may not correspond to the working capital an efficient operator would need For example the adjustments to operating cost would affect the level of working capital by altering the level of creditors Further the efficient operator may have a smaller level of working capital due to better management of debtors creditors cash and the supply chain However given that the cost of working capital is likely to be small in relation to other costs such inefficiencies might not be material 3 4 3 1 How the working capital is used in the model The total Working Capital WC is determined using the method outlined above Next the total WC value is multiplied by the cost of capital to find the actual cost of WC The cost of WC is then calculated as a fraction of the total costs sum of costs of Core Access and Co location This is then converted into a percentage value that is used to uplift the cost of services The percentage value used in the model also includes the effect of common business costs as well as WC Date 30 11 2009 Page 28 Worki
41. systems In order to increase model transparency and avoid unnecessary complexity these cost categories are allocated according to spacing requirements m Security guard security system site preparation and site maintenance are thereby allocated in the same fashion as in the previous version of the hybrid model Costs that relate to trenching ducting fibre and IP network equipment are specified with a wide range of types This enables the costs to be determined based on a realistic profile of the different elements required For example trenching costs vary by the surface type therefore different unit costs are specified for each surface Also there is a profile determining the likely amount of each type of trenching required 9 8 3 Cost allocation methods 9 8 3 1 Cost allocation between Core and Access The allocation of costs from the core network model to access products takes place in one of two ways For active equipment and the MDFs the costs are allocated to access related network elements and then the cost of these elements is allocated to the relevant access products in the consolidation model For trench duct and fibre trench sharing parameters defined in worksheet I Technical are applied to the costs at the bottom of worksheets C_Trenching Transit and C_Trenching Local in order to exclude from the core model costs not related to core products The total amounts for shared trenching also feed through to the access mode
42. the cost of preparing indoor co location space for racks and cables Rack space includes two cost categories gt Room fit out including costs for floor lighting fire alarm and planning gt Cable trenches including costs for the actual trenches cabling and planning The costs for the two categories are shown per rack space In previous versions of the model rack space was based on a standard size of 1 5 m However in practice in technical houses there is not sufficient space for an additional rack TDC has however indicated that there will possibly be space in TDC s racks that the other operators will be able to use Therefore a new product rack space in a technical house has been added This product is defined as 1 5 of a standard half depth ETSI rack 600x300x2200 mm Two charges have been included gt A one off charge covering 1 5 of room fitting cable trenching as in the existing normal co location product installation of co location gt A recurring charge covering 1 5 of space for co location assuming that a rack 600x600x2200 mm covers 1 5 m2 a rack of half the size 600x300x2200 mm is assumed to cover 0 75 m2 1 5 of an ETSI standard rack 600x300x2200 mm For room fit out the cost of lighting is included in the common site costs The ongoing operating expenses for air conditioning are also included in the common site costs Finally the cost of fire alarms is modelled separately but as part
43. the primary access network Finally information is also provided on the length of the final drop to the subscriber premises 8 2 2 Sub Loops In 2006 the Hybrid model was extended to include the modelling of sub loops within the Access network as described in Chapter 5 above NITA is conscious that moves towards the incorporation of next generation access NGA technologies and techniques are likely to fuel demand for such sub loops One possible consequence of this might be that the currently defined point of access a nationally averaged primary distribution point or PDP might not necessarily continue to be best representative of where alternate operators actually require access Date 30 11 2009 Page 58 The impact of a move towards Next Generation Access on the existing model could occur in a number of ways in particular gt A potential need to geographically de average the PDP locations fed by fibre gt A potential need to model access backhaul services such as lit dark fibre or ducts from the PDP back to the exchange or technical house gt A potential need to model full fibre local loops The model now allows the user to specify the number of PDP locations fed by fibre referred to in the model as Active Street Cabinets for each of the four geotypes This geotype disaggregation is fed all the way through to the output sheet of the consolidation model producing cost based prices for sub loops by geotype in addition to the
44. the revised Hybrid Model are shown in Table 3 9 Table 3 9 Building Values by Geotype 2 1 Valuation incl technical houses LandB Average Valuation DKK m2 10 228 4 849 3 624 3 514 7 210 Hereof value of land DKK m2 SSSI 2 718 401 124 83 557 Building values DKK m2 7 509 4 449 3 499 3 431 6 653 Ratio land m2 building m2 SSS 0 55 2 51 6 90 9 26 2 19 2 2 Valuation per geotype excl technical houses Stor By LandA Land B Average Valuation DKK m2 10 249 4 898 3 511 3 699 7 310 Hereof value of land DKK m2 a 2 728 428 162 84 650 Building values DKK m2 __ 7 520 4 470 3 349 3 615 6 660 Ratio land m2 building m2 _ 0 55 2 34 5 76 9 70 1 90 Date 30 11 2009 Page 31 This chapter sets out to describe the assumptions and methods used for converting Gross Replacement Costs GRCs derived in the Hybrid Model to annual capital charges annualised costs The major assumptions affecting the annualised costs are gt Cost of Capital CoC gt Asset lives gt Price trends gt Output trends 4 1 Annualisation methodologies 4 1 1 The FCM principle The Financial Capital Maintenance FCM principle is used when the annualised costs for assets in the Hybrid Model are estimated The concern of FCM is to maintain the financial capital of the company This maintenance is achieved when the value of shareholder funds is the same in real terms at the start and end of the period Using the FCM method implies adding
45. to describe the methodology and assumptions behind the access model and describes the changes that have been made in the revised Hybrid Model to reflect the increasing importance of sub loops and DSLAMs at the street cabinet resulting from the transition to Next Generation Access NGA 8 1 The Access Network in Denmark The access network connects the end user with central equipment in the core network Subsequently the related costs primarily depend on the number of subscribers and less so on the volume of traffic Consistent with this access can alternatively be defined as those services which allow the subscriber to send and receive traffic Both definitions suggest that the access network includes all costs for cable and trenching associated with subscriber lines between the customer premises and the concentrator These definitions also suggest that the access network includes analogous costs for other lines such as those for public call boxes between the customer premises and the concentrator Furthermore the definitions suggest that the access network includes the line card within the concentrator DSLAM and potentially other subscriber related costs This is consistent with the first definition since line card requirements are driven by the number of subscribers or more accurately by the subscriber requirements for lines It is also consistent with the second definition since the line card is essential to sending and receiving traffic H
46. total costs they are still important for the final costs These costs are allocated exclusively to switched interconnection services and wholesale access services The costs are summed up as a total and then allocated between the two service categories using a distribution scale as shown in Table 3 6 Table 3 6 Allocation of Interconnection Specific Costs Regulated access Non regulated 5 4 Specific IC and Commercial Costs services Regulated core services Co location services _ wholesale services Split between access and interconnect 20 40 20 20 Regulated core split between voice and data eg BSA Voice Data 70 30 Source Consolidation model I FA_Costs The costs are allocated to the services using a multiplicative mark up The mark up is calculated from the interconnection specific costs for the increment in relation to the overall operating costs for the increment Date 30 11 2009 Page 26 3 4 2 3 Allocation of non network related costs The LRAIC cost base also contains non network related costs that is costs which are necessary for running a telephone company but which may not be allocated immediately to the individual network elements These elements are allocated directly to the services by use of a mark up method The mark ups used in the hybrid model are shown in Table 3 7 Table 3 7 Mark up for Other Annual Indirect Common Costs incl Operational Capital Mark up pct Services in the acce
47. up the Hybrid Model a user should understand the purpose of each workbook and should have at least a high level understanding of the structure and interdependencies of the different models Note that a user need not have detailed Excel skills but a basic knowledge is expected If the intention is to modify the models or to investigate the algorithms used then some advanced Excel skills may be required to understand the details of how data is processed 7 1 1 Description of model package The models have been developed using a standard software package Microsoft Excel However total functionality cannot be guaranteed for users with a version older than Microsoft Excel 2003 7 1 2 Quick start and how to find main results All results are in the consolidation model It is not necessary to have the other models open to see the results When opening the consolidation model Excel will offer the choice of updating links If the model is opened for the first time this is not necessary Links should be updated however if changes have been made to the other models If you update links from the consolidation model then the other model files must be available in the same directory When the model is opened a standard precautionary message may appear containing a warning on the presence of macros in the model If this is the case and if the file is from a trusted source users should click on the button Enable Macros otherwise the macro
48. where assumptions about the manpower consumption for different tasks can be entered Horizontally the sheet has the different staff types administrative technician and academic A weights column is used to allocate shared activities such as end user visits or transport time between services that are completed concurrently so as to avoid double counting This is explained in more detail in Section 10 2 2 of this report I Demand This sheet contains the various demand assumptions for co location services Where no demand data can be obtained as for new services a dummy value of 1 has been used to avoid products being excluded from the model and to ensure they receive an allocation of common costs as appropriate Date 30 11 2009 Page 141 14 3 2 Calculation sheets There is one calculation sheet C_Costs This sheet contains calculation and allocation of costs to services using the information in the input sheets For each product it shows the various parameters such as installation costs maintenance price trends lifetimes and opex trends which will be used in the consolidation model to calculate annualised service costs where appropriate 14 3 3 Output sheets O_Consolidation This sheet contains cost at cost category level and demand at service level which both are output to the Consolidation model Information in this sheet is fed in from the C_Costs sheet The calculation of final service costs after addition of overh
49. 0 Date 30 11 2009 Page 68 and and is an input in the model An extract from the model for the MDF Slangerup is shown below Figure 8 3 Extract from the Hybrid model showing the geometry for PDP and SDP zones Diagram 2 Locations and geometry for PDP and SDP zone Note each grid square is 0 25km x 0 25km Excel scaling may distort the image I _ ASM Switch for Zone no 574 Blahgj Focal point demand weightet centroid for ASM zone PDP 2 zone boundary PDP2 cage A Centre of SDP zone 1 2 3 172 SDP zones in the ASM zone 0 e fConnections between the centers of ASM PDP SDP zones ICable connections ASM PDP SDP A SDP2 zone boundary assuming itis roughly street shaped More specifically the figure shows that gt The PDP location strategy gives the freedom to locate the PDP either at the edge of the PDP area the brown ring if a value of 1 has been selected or closer to the subscriber if a value lower than 1 has been selected gt The SDP location strategy allows an SDP to be located at the border of the SDP zone nearest to the PDP if a value of 1 is selected or reciprocally closest to the subscriber if a value of 0 is selected The figure also shows the prescribed shape of the PDP and SDP areas By default the PDP areas are circular and the SDP areas rectangular However the model undertakes two sanity checks which may imply that the SDP
50. 002 The copper requirements for the whole network are worked out through the means of the weights attributed to each zone The 20 sample zones are used to calculate the total amount of copper cable necessary to connect all of the actual and prospective customer sites available These totals are then adjusted downwards via scaling factors to take account of the actual number of customers connected The number and the size of modelled SDP and PDP cabinets depends on the number and the size of the modelled SDPs and PDPs for each of the twenty sample zones again the requirements for the whole network are worked out through the means of the weights attributed to each zone The number of modelled SDPs per zone is worked out as the ratio between number of subscriber lines served in each zone and connected copper pairs per SDP an input of the model The number of modelled PDP per zone is an input of the model The model now also includes the facility to model active street cabinets located at PDP sites by geotype Note that where active street cabinets are assumed to occur then care needs to be taken in deciding the average number of subscriber lines covered by each cabinet location since too small a number will result in a location with insufficient subscribers to warrant an upgrade and too large a number will negate the improvements of shortening the copper loop length In general a suitable number of potential customers per active street cabin
51. 12 Rural B DKK 375 DKK 6 10 1 3 2 Common site cost Common site costs are calculated in the core model and allocated to co location in the consolidation model Therefore these costs are not shown in the actual co location model Common site costs include the following cost categories related to exchange sites gt Power supply gt Air conditioning gt Security system gt Site preparation gt Site maintenance gt Security guard Each cost category contains the calculated annual capital expenses depreciation and cost of capital for equipment materials and installation plus the annual operating expenses for maintenance buildings and power consumption As an example the cost of air conditioning including annual power consumption related to co location space is included in the site cost category of air conditioning unit For the allocation of common site costs the following allocation keys are used Date 30 11 2009 Page 107 gt Power supply unit average power requirement kWh gt Air conditioning unit average power requirement per m kWh m per exchange type gt Security guard security system site preparation and site maintenance space requirement m Please note that costs of power supply are not allocated to the common site costs for co location The costs for power supply are modelled separately in the co location model 10 1 3 3 Rack space The price of rack space covers
52. 2 4 1 3 Standard annuities 33 4 1 4 Tilted annuities 33 4 1 5 Economic depreciation 33 4 2 Annualisation methods used in the Hybrid Model 34 4 3 Calculation of economic depreciation 34 4 4 The Cost of Capital 38 4 5 Asset lives 38 4 6 Price trends 40 Date 30 11 2009 Page 3 4 6 1 Trench and duct in access and primary networks 41 4 6 2 Copper cables 42 4 6 3 Storage and factory buildings and land 43 5 Additional Products Added in 2006 44 trae G e a 5 1 Introduction We Ps phere eee ong ee ts ok 5 2 Impact on access and core models 44 5 3 Sub loops 45 5 4 Bitstream Access 46 5 5 Ancillary products 49 6 Product Changes in 2009 50 6 1 Introduction 50 6 2 Main LRAIC service costs 50 6 3 Other LRAIC service costs 50 6 4 Co location services 51 7 How to Use the Revised Hybrid Model 52 7 1 Introduction 52 7 1 1 Description of model package 52 7 1 2 Quick start and how to find main results 52 7 2 Overview of common model features 52 7 3 How the models are linked 55 7 3 1 Important actions 56 8 The Access Model 57 8 1 The Access Network in Denmark 57 8 2 Overview of the Hybrid Access Model 58 8 2 1 General Approach 58 8 2 2 Sub Loops 58 8 2 3 Key network elements 59 8 3 Main definitions and assumptions in Access 61 8 4 Modelling the access network 61 8 4 1 Trench and duct 61 8 4 2 Mini duct 64 8 4 3 Copper and fibre equipment 66 8 4 4 Street cabinets 70 8 4 5 Fibre access network 74 8 5 Treatment of shared costs in the Access
53. 2 Edge Min TDM Gateway 3 Spare1 1 All Min TDM Gateway 4 Spare2 1 All Min IMs 1_ Session Border Controller 1 Distribution Min IMS 1 Session Border Controller 2 Edge Min IMS 0 Spare 0 All Min IMS 2 IMS Core 1 Call Related Call IMS 2 IMS Core 2 Subscriber Related Subs IMS 2 IMS Core 3 Other Subs IMS 0 Spare 0 All Subs IMS 3 Application Service VolP 1 Subscriber Related Subs IMS 3 Application Service VolP 2 Other Subs 9 5 Core trenching Due to the increasing importance of the cost of trenching duct fibre on the resultant product costs the model now adopts a radical new method of assessing the requirement for core trench km A study was conducted of the core trench routes required to interconnect the various core sites within TDC s network in Denmark The Date 30 11 2009 Page 91 study was based around a dataset of the node locations provided by TDC and utilised Microsoft Autoroute software The analysis was conducted in two separate stages one considering the trench km necessary to connect all of the higher level nodes from Layer 3 Edge routers upwards together and the second to connect all of the DSLAM MSAN sites to the Edge router sites In both cases the analysis covered the complete set of sites and also included additional trench km to provide for a reasonable amount of trench rings so as to provide physically resilient routing possibilities For the higher level nodes each route segment connection between two node
54. 3 1 Input sheets 141 14 3 2 Calculation sheets 142 14 3 3 Output sheets 142 14 4 Description of worksheets in the Consolidation Model 142 14 4 1 Utility sheets 142 14 4 2 Input sheets 143 14 4 3 Calculation sheets 144 14 4 4 Output sheets 145 15 Appendix D Bilag vedr kortlagte andre tjenester for LRAIC 2010 146 Date 30 11 2009 Page 6 1 1 Background In 2002 the National IT and Telecom Agency NITA commenced regulation of a number of TDC s wholesale prices in accordance with the LRAIC pricing methodology To facilitate this process an Excel based model was constructed in collaboration with the industry including TDC with the intention of incorporating suitable inputs both from a Top Down as well as a Bottom Up perspective This so called Hybrid model has since been updated annually In addition to the annual updates to the model NITA is committed to consider every three years whether there are suitable grounds for a more extensive update Since the last major review in 2005 the pace of change and progress in the telecommunications market has continued unabated In particular gt The trend of decreasing prices for the conveyance of traffic both voice and non voice shows no sign of stopping even though there is now a declining volume of traditional telephony traffic being carried over the legacy circuit switched network gt New services such as IPTV are progressively being introduced that are centred on the
55. GE Modules Mpps 5 0 1GE ports Gbps 5 0 Layer 3 Distribution Routers Configuration Rules Minimum number of chassis per site Maximum number of modules per chassis Maximum number of 10GE per module Maximum number of 10GE per chassis 16 Usage Drivers Allowance for Spares Chassis Mpps 1 0 Modules Mpps 5 0 10GE ports Gbps 5 0 Layer 3 Core Routers Configuration Rules Minimum number of chassis per site Maximum number of modules per chassis 8l Maximum number of 10GE per module as Maximum number of 10GE per chassis 40 Usage Drivers Allowance for Spares Chassis Mpps 1 0 40GE Modules Mpps 5 0 40GE ports Gbps 5 0 10GE Modules Mpps 5 0 10GE ports Gbps 5 0 DWDM Equipment Configuration Rules Base Units per link a2 Maximum unamplified distance 100 km Usage Drivers Allowance for Spares Base Units Gbps 1 0 Amplifiers Gbps 1 0 Lambda inserts Gbps 5 0 Date 30 11 2009 Page 133 IP TDM Media Gateways Configuration Rules Minimum number of chassis per site A per Distribution Layer site Include OLO facing line cards in Core Model costs No 13 Sites Media Gateway Controller A per 150 000 BHCA Distribution Sites Media Gateway Chassis 1 per site Media Gateway Front Card o dTpa BH Erlangs Media Gateway Back Card diper BH Erlangs Edge Sites Media Gateway Small diper 22 IBH Erlangs Media Gateway Medium o dpe 44 BH Erlangs Media Gateway Large A per 132 BH Erlangs Usage Drivers
56. GIS Trench and Duct and then used to derive the trench and duct information for each sample area split between ASM PDP and PDP NTP C_Trench and Duct This is based on inputs from Z_G S Trench and Duct and calculates the total trench and duct kilometres by road type for each of the four geotypes This is then broken down further by terrain type for trench and duct type C_Share_Routes This spreadsheet works out the amount of shared trench and duct to be allocated to the core and access network on the basis of the inputs in Z Share Routes in terms of both length of common trench and allocation key to the different increments and the calculations performed in C_Trench and Duct that estimate the total amount of trench and duct in the access network C_Cables and Nodes This is by far the most complicated spreadsheet in the whole model However most of the calculations performed in this spreadsheet are aimed at estimating the length of the loop by geotype and part of the network It is important to bear in mind that most of the calculations shown in this spreadsheet refer to the sample area identified in cell J3 All information is then aggregated by multiplying the results of each sample zone by the weights attributed to each zone in J_GIS Zones in order to derive results for the whole network Date 30 11 2009 Page 137 C_Resources This spreadsheet lists all the results in terms of capital requirements and operating costs for each geotyp
57. I FA Costs and C_FA Costs The first worksheet of these I FA Costs contains five basic sections Functional area inputs Staff costs Calculation of pay costs Calculation of total costs Overheads Que NO Date 30 11 2009 Page 121 11 2 1 Functional Area inputs The functional area inputs consist of six sub sections The first of these provides a set of operating cost drivers that are combined with the second sub section to derive the numbers of FTE per functional area These operating cost drivers include such measures as the number of core nodes the number of km of core trench and the number of DSLAM lines The measures essentially extract relevant data from elsewhere in the model and as such could relatively easily be expanded as and when necessary The second sub section consists of a set of input parameters related to ongoing and one off staffing requirements for each functional area Each functional area has four inputs associated with it gt A minimum number of FTE for that area gt An additional number of FTE that are required as the relevant driver increases in size gt The increase in size of the driver that relates to the additional number of FTE gt The actual driver itself such as km Core Trench This sub section also allows the user to input a different set of staffing assumptions that can override the calculations based on the parameters and also to include an efficiency improvement that will be a
58. IC Hybrid Model The FA method identifies a number of major cost categories that are based on the basic functional areas of a telecommunications provider as shown in Table 3 2 This approach to the modelling of operating expenditures is comparable to the method normally used by providers developing business plans LRAIC operating overhead and annualised indirect capital costs or FA costs should thus comprise the costs that are relevant for the wholesale provider in the core and access increments Any cost being directly related to end user activities should be excluded This ensures that costs for other increments not modelled directly in the Hybrid Model are excluded from the model Moreover costs that are modelled indirectly e g non PSTN costs are also excluded from the final allocation of costs to specific interconnection services The method is based on two main steps gt Statement of costs per functional area gt Allocation of costs These steps are described in detail below 3 4 1 Statement of costs per Functional Area The cost categories are defined in accordance with the overview in Table 3 2 Date 30 11 2009 Page 21 Table 3 2 Detailed Summary of the Functional Areas Costs Network related on going costs Network related one off costs Non network related cost Interconnection specific and commercial operating costs easy IMS management and planning IMS maintenance Network Manag
59. It also summarises in Section 6 the accommodation related overheads for different types of equipment such as DSLAMs MSANs Edge Routers or Core Routers based on inputs from Z Costs I Technical This sheet contains the various input information necessary to size the equipment used in the core network Date 30 11 2009 Page 139 I_Trenching_Transit This worksheet contains the inputs related to higher level trenching in the core network and also is used to calculate the necessary fibres and fibre link speeds usage for the links at those levels I_DWDM_ graphic This is a non functional that is graphic only worksheet to provide an illustration of the potential DWDM network considered in worksheet 7 DWDM I DWDM This worksheet allows the user to determine the extent of the modelled DWDM network I Trenching Local This worksheet contains the inputs related to lower level trenching in the core network that is from the Edge Router downwards I Node _ Equipment This worksheet is used to calculate the required node equipment at the DSLAM MSAN sites in the core network Each site is treated individually I Costs This worksheet contains inputs for the per unit cost of transmission infrastructure and equipment and common site costs The sheet also contains inputs for economic depreciation parameters 14 2 3 Calculation sheets C_Equipment This worksheet is used to calculate and summarise the costs of all equipment addressed
60. Media Gateway Controller Call Media Gateway Chassis Min Media Gateway Front Card Min Media Gateway Back Card Min IMS Configuration Rules 1 per minimum of 1 per Each IMS contains Session Border Controller A per SBC at Edge location IMS Core Softswitching Aper RTU licences 10 000 per 10 000 per Fixed per IMS Ape Application Service VolP 2 per Application Service Centrex 4 per Proportion of business subscribers using Centrex 10 Usage Drivers Session Border Controller Min Softswitching Call RTU licences Subs related Subs RTU licences Call related Call Fixed per IMS Subs Application Service VoIP Subs Application Service Centrex Subs Date 30 11 2009 Allowance for Spares 10 10 5 0 5 0 200 000 Voice subscribers Distribution Layer site 13 Sites ____ 2 000 BH Erlangs No 425 000 JBHCA 10 000 Residential subscribers 10 000 Business subscribers IMs 50 000 Voice subscribers 50 000 Business subscribers using it Allowance for Spares 10 1 0 10 10 10 Page 134 14 1 Description of worksheets in the Access Model 14 1 1 Input sheets I Interface This sheet contains inputs from other parts of the model that are shared with Access via Excel links It provides information about the original input source to allow the user to locate the yellow or blue modifiable input cell I Network_Demand This spreadsheet is split into three sections The first and thir
61. Report on the LRAIC Model and User Guide Revised Hybrid Model version 3 1 IT og Telestyrelsen November 2009 Published by IT og Telestyrelsen Holsteinsgade 63 2100 Copenhagen Phone 45 354500 00 Telefax 45 3545 0010 E Mail itst itst dk www itst dk Table of Contents gt 1 Introduction 7 1 1 Background 7 1 2 The use of LRAIC price estimation methods 8 1 3 The Hybrid Model 8 1 4 The purpose of this report 9 2 Overview of Methodology Used to Build the Hybrid Model 10 2 1 Defining the increment 10 2 1 1 Shared and common costs 11 2 1 2 Defining the core and access increments 12 2 1 3 Other increments 14 2 2 Other main assumptions 14 2 2 1 The scorched node assumption 14 2 2 2 A Network Element approach 15 2 3 Services in the LRAIC Hybrid Model 15 2 3 1 Voice and broadband data services 15 2 3 2 Leased lines 16 2 3 3 Other services 16 3 Costs 17 3 1 Defining costs 17 3 2 The treatment of common costs 18 3 3 Unit costs 19 3 3 1 Sources of cost estimates 19 3 3 2 Cost estimates in the Hybrid Model 19 3 3 3 Data input and confidentiality 20 3 4 Operating costs 21 3 4 1 Statement of costs per Functional Area 21 3 4 2 Allocation of costs 25 3 4 3 Working capital 28 3 5 Building costs for exchanges and technical houses 29 3 5 1 Calculation of the real property costs 29 4 Annualisation Assumptions 32 4 1 Annualisation methodologies 32 4 1 1 The FCM principle 32 4 1 2 Straight line depreciation 3
62. The service is only modelled as the standard service within normal working hours and is thus calculated on the basis of salary costs for normal working hours The hybrid model also includes costs for supervised access for co location The cost is the hourly rate for a technician plus a mark up for common cost Date 30 11 2009 Page 118 10 3 How to use the Co location Model 10 3 1 Model scope The co location model calculates the resources required to offer co location services in the fixed network at exchanges or technical houses The main components of this are gt Site costs gt Cabling costs gt Power cooling and ventilation After calculating the resources needed to offer these services the model uses the results to calculate the capital cost of those resources and the annual operating costs of such services The Co location model further contains a range of products which do not lend themselves to the modelling approach used in the Core and Access models These include services related to installation of end user equipment order processing and the provision of raw or shared copper whose costs derive largely from the use of manpower These other services are discussed in detail in Section 10 2 above 10 3 2 Model structure For ease of navigation the Co location model includes a map of all major information flows between them This is reproduced in the diagram below Figure 10 2 Co location Model Navigation Map I
63. a holding gain or loss to the annualised cost when the price of the asset changes during the course of the year A number of annualisation methods may be used in combination with the FCM principle These include gt Straight line depreciation gt Standard annuity gt Tilted annuity gt Economic depreciation The following sections give a brief outline of each method To ensure maximum flexibility the Hybrid Model allows users to apply any of these four annualisation methods to individual asset classes or to the model as a whole The method preferred by NITA the default setting in the model complies with the principles set out in the MRP and is described in Section 4 2 below 4 1 2 Straight line depreciation Straight line depreciation is the method most commonly used in financial accounts It simply spreads the original cost of an asset evenly across its economic lifetime so that an asset costing 5000 would be recovered over 5 years by charging 1000 to depreciation each year The method is popular because of its simplicity but is criticised for not reflecting economic reality as assets typically lose most of their Date 30 11 2009 Page 32 value in the early years It also ignores the cost of capital which must be calculated separately 4 1 3 Standard annuities Annuity methods also spread the cost of an asset over its economic life but in addition take account of the opportunity cost of capital ie the interest forgone whi
64. ad to be made for the different levels of source data available The previous data set for Q2 2004 Q1 2005 contained sales volume and growth data at municipality or city level From Q2 2005 data were only available at Amt level and from Q1 2006 data were available at Landsdel level Therefore the data were grouped into three distinct periods gt Period 1 Q2 2004 Q1 2005 gt Period 2 Q2 2005 Q4 2005 gt Period 3 Q1 2006 Q2 2008 A correction factor was calculated for each period by the following methods gt Period 1 Obtained from the 2006 model for each municipality and averaged for new Landsdele gt Period 2 Projected for each municipality based on 2005 average growth per Amt The data were then averaged for the new Landsdele as for Period 1 Date 30 11 2009 Page 30 gt Period 3 A correction factor was calculated by the same method as for the original analysis using Landsdel level data from Denmark Statistics To derive the final continuous correction factor for Q2 2004 Q2 2008 the correction factors for the three periods were averaged weighted by the number of quarters in each period Finally the new correction factor was applied to the TDC data to derive updated building values This was done by applying the average Landsdel correction factor to each property located within the Landsdel using an updated postcode lookup table obtained from the Danish Post Office The building values used in
65. age including employer or employee paid pension for each of these categories Estimates for gross wages are based on Date 30 11 2009 Page 22 publicly available data drawn on the 1st of October 2005 adjusted to reflect 2009 costs Data from the Danish Metalworkers Union Dansk Metal was used for the technicians and for the academic employees data was collected at The Association of Danish lawyers and Economists Danmarks Jurist og konomforbund and The Society of Danish Engineers Ingeniorforeningen i Danmark On the basis of these estimates an effective hourly wage for each type of position was calculated Assumptions have been made regarding the number of possible working days per year weekly working hours and corrections for days off and other absences including illness training courses and general administrative activities Furthermore extra costs such as higher wages for managers have been taken into account by effectively moving 10 of technical staff into the higher academic wage bracket The input parameters are shown in Table 3 3 below Table 3 3 Staff Cost Calculations 5 1 Working days assumptions Adm staff Potential working days per year Days i 252 Absence sickness injury other Days 10 Holidays Days Ez 30 Administration Days E 10 Courses Days a 3 Actual number of working days per year Days 199 0 Average number of hours per day for administrativr staff Hours 7 4 Average number of
66. al number of SDPs in the network 431 551 Table 8 8 Average number of connected copper pairs per PDP Geotype 1 477 Geotype 2 475 Geotype 3 123 Geotype 4 52 Total network 173 Table 8 9 Average number of connected copper pairs per SDP Geotype 1 17 Geotype 2 9 Geotype 3 5 Geotype 4 3 Total network 5 7 Different assumptions regarding the number of cabinets have an impact on the loop lengths The more cabinets in the access network the shorter the total loop length will be This is because the presence of more cabinets reduces doubling back with lots of SDPs you take subscribers directly to their homes with few you will often go further and then double back On the other hand more cabinets should increase the length between the MDF and the SDP increasing the number of SDPs takes you closer to the subscriber and decrease the distance between the SDP and the NTP The method used by the model when calculating distances between the connection points in the access network considers assumptions on the number of cabinets 8 4 4 1 Active Street Cabinets Version 2 4 of the model released in 2006 introduced the concept of Active Street Cabinets essentially street cabinets housing small DSLAMs and intended to reduce the effective length of the local loop and hence the maximum speeds achievable with xDSL It was agreed at that stage that such cabinets should be assumed to be located at th
67. apters 8 and 9 respectively The LRAIC methodology is based on these main increments The incremental costs of the core increment are those incurred when adding a core network when the access network is already present similarly the incremental cost of an access network are the costs incurred when adding an access network when the core network is already in place The LRAIC of co location is the cost incurred when Date 30 11 2009 Page 10 providing co location services These definitions include the services provided by the SMP operator s network division to its own retail division as well as services provided to other operators The unit costs typically increase following the increment size This is due to the large proportion of the operator s cost base that changes less than proportionally with changes in output If your current output is different from zero the marginal duct costs will be very small for almost all changes in the output For this reason the unit costs will be largest in the third definition of the increment and smallest in the first definition Unit costs in the second definition will be closer to the first than the third definition The definition of the increment used for LRAIC means that fixed costs specific to either the core or access networks are included These costs include a large proportion of optical fibre costs and the trenching and duct costs in either the access or core networks but not the trenching a
68. ated on the basis of an analysis of 20 sample areas these were then aggregated to a national equivalent using a series of weights The method of calculating the cabling made it relatively straight forward to estimate the cabling costs of sub loops since the analysis separately considered ASM PDP links from PDP SDP and SDP NTP links For the trenching it was concluded that it would be necessary to assess the amount of trenching required for each of the 20 sample areas contained within the existing model in order to then more easily distinguish between ASM PDP trenching and PDP NTP trenching This involved adding a separate worksheet to the Access model that instead of starting with an analysis of the road system by geotype contained an analysis by sample area A similar process was then followed to that contained within the existing C_Trench and Duct worksheet to arrive at tables detailing the required amount of trench and ducting and in addition distinguishing between primary ASM PDP and secondary network PDP NTP In recognition of the fact that the 20 sample areas were unlikely to yield the previously calculated national totals for trench and duct the weights were originally designed to match a different set of national figures the additional worksheet normalised the sample by sample analysis such that identical total amounts were calculated when the sample weights were applied 5 4 Bitstream Access Bitstream Access is the term used
69. ated with these streets in the original model is very low as if these roads had no sites at their sides NITA has identified through the use of the DAV database the location of all Ekstra vej in Denmark in order to verify whether they were included in the INCA maps provided by TDC Unfortunately only two of these were part of the INCA maps provided and therefore it has not been possible to meaningfully estimate a separate conversion factor for these types of road However they are included in the overall database along with squares for a meaningful measurement of which concerns were also raised and the conversion factors for most of the road length in Denmark have been estimated making reference to a meaningful sample of data The table below shows NITA s conversion factors as estimated through analysis of INCA maps and the econometric method Table 8 3 Overview of conversion factor estimates Road type The Hybrid model AA 1 80 AB BA 1 61 AC CA 1 40 AD DA 1 10 BB 1 63 BC CB 1 18 BD DB 0 89 CcC 1 00 CD DC 0 70 DD 0 40 Table 8 3 shows the conversion factors used in the hybrid model The total length of trench is hereby estimated to 124 245 kilometres 10 instead of 16 classes like AB and BA would count as a single road class Date 30 11 2009 Page 63 8 4 2 Mini duct The hybrid model estimates mini duct in the following way gt by measuring the number of subscriber sites in each o
70. ature of the core network remains but the type of equipment located at each level of the hierarchy differs In principle the hierarchy will consist of from top downwards gt Core Layer 3 IP Routers gt Edge Layer 3 IP Routers gt Layer 2 managed Ethernet traffic aggregation switches gt DSLAMs MSANSs containing a mix of xDSL and POTS line cards In addition to the above in order to provide a voice telephony service there will be a need to overlay certain other equipment including Date 30 11 2009 Page 79 gt Softswitches and associated servers to control voice telephony call routing and call data record management gt Media Gateways to provide a TDM based level of interconnection whilst the alternative operators themselves still maintain and operate circuit switched networks IP based equipment tends to be cheaper than TDM based equipment when comparing similar levels of throughput This means that a greater proportion of asset related costs will now be due to the trenching ducting and fibre optic cables However it is still necessary to assess the configurations of the equipment upon which the model is based and to what degree the necessary investment ought to vary with service demand Thus the model still needs to incorporate the ability to alter the configuration inputs for each major type of equipment in order that the level of investment within the core network can adequately flex with traffic demand 9 1 2
71. broadband that is packet switched network These services also cover displacement services such as VoIP which are acting as an additional driver for the decline in use of equivalent services carried over the legacy network gt The development of the packet switched network has reached the stage where it is now commonplace for industry observers to reflect on the need for a final migration of all remaining circuit switched services to an all IP core network resulting in the so called Next Generation Network NGN gt The demand for increased bandwidth to the end subscriber is also driving deployment of advanced technologies and techniques into the access network resulting in the deployment of so called Next Generation Access NGA networks These typically revolve around the deployment of fibre into the access network either as a means of reducing the overall length of the final copper connection or as a means of displacing it entirely with Fibre to the Home Building FTTH FTTB During 2008 NITA concluded a consultation process where it elicited the views of the industry participants regarding the need for a more extensive revision of the current Hybrid model As part of this process NITA impressed on the industry the implications of the use of an LRAIC pricing methodology namely that it required the modelling of an efficiently operated company making use of an optimally structured network infrastructure based on latest genera
72. by the core model C_Route Table This worksheet takes feeds from I Route Table and applies the routing factor inputs to traffic volumes measured in Busy Hour Erlangs Busy Hour Calls Busy Hour Gigabits per second Busy Hour Mega packets per second or number of Subscribers For voice traffic it applies the uplifts for call holding time and unsuccessful calls specified in 1_Product_Demand There are separate calculations for required capacity and actual capacity Actual values refer to Year 0 data inputs from I Product Demand which are used to calculate unit service costs Values to use or required capacity are applied to projected traffic volumes for year 1 2 3 or 5 and are used for network dimensioning C_Trenching_Transit This worksheet is used to calculate trench and trench related costs for the higher levels of the core network C_Trenching Local This worksheet is used to calculate trench and trench related costs for the lowest level of the core network that is from the Edge Router downwards Date 30 11 2009 Page 140 14 2 4 Output sheets O_Consolidation This worksheet collects the cost for each cost category and then uses a look up technique to pick up other parameters that relate to that cost category These include equipment installation costs scrap values lifetimes price trends maintenance costs and various parameters used for the calculation of economic depreciation and other annualisation methods The
73. ccess and core models is that the co location model does not model direct network related costs Instead the model estimates a range of indirect network costs and bundles them to form different co location products and services The modelled co location services are gt Co location space indoor m and outdoor m gt Rack space gt Power supply 48 VDC 16 A 48 VDC 30 A 48 VDC 60 A and 230 VAC 10 A gt Cables 8 24 100 and 300 pairs and in span 10 1 2 1 Types of co location services and costs The most important co location services for all access seekers are indoor co location space and rack space Power supply and cables can be seen as additional services For power supply there are different choices and combinations of 48 VDC and 230 VDC connected to the power supply of the available buildings It is also possible for access seekers to order their own power supply and meters from the utility companies Similarly for cabling the choices are between different possibilities depending on the access seeker s demand for capacity and on whether the co location is used for interconnection or raw copper Another difference between the two main co location services and the additional services is that a lot of the cost related to the main services e g space is calculated in coherence with similar demands in other parts of the hybrid model Conversely the additional services have very few cost categories a
74. ch to gaining an understanding of the model is to work through each sheet in turn starting with inputs moving to calculations and then outputs One way to improve understanding is to edit the input data recalculate and investigate the effect of the changes on the results These can be easily checked in the sensitivity analysis sheet in the Consolidation model Date 30 11 2009 Page 102 The final costs for all modelled core products are shown in the output and results sheets of the Consolidation model A sheet by sheet description of the Core model is provided in Appendix C of this document Date 30 11 2009 Page 103 This chapter sets out to describe the general characteristics of the co location model and the modelling of other services in the hybrid model Other services cover the services regulated by LRAIC which do not directly use the access or core network or which are related to specific work processes 10 1 Co location services 10 1 1 Definition The Executive Order Number 930 Section 1 1 refers to co location as the sharing of facilities such as buildings exchange equipment etc and includes co location as an interconnection product linked to the three other interconnection products namely 1 exchange of traffic 2 lease of infrastructure capacity and 3 service provider access Section 1 7 of the Executive Order defines sharing of facilities such as buildings exchange equipment etc as gt Acces
75. ch would have been earned had the cash been invested elsewhere Therefore annuities consist of two separate elements the annualised cost of the asset depreciation and a financing or cost of capital charge In a standard annuity the annual charge remains constant over the life of the asset Again the method has been criticised for failing to reflect the true depreciation profile of the asset and for assuming constant asset prices Standard annuity formula Annuity Purchase Price WACC 1 1 0 WACC 4 1 4 Tilted annuities Tilted annuities relax the assumption of constant prices In fixed networks asset prices tend to fall over time whereas infrastructure costs digging trenches for example tend to rise over time If for example the annualisation method ignored falling prices Entrant 2 would have an advantage over Entrant 1 as it would benefit from lower asset prices and consequently lower depreciation charges When asset prices are falling a tilted annuity recovers more of the capital value in the early years and vice versa which ensures that two entrants with an identical asset base though acquired in different periods have identical depreciation charges Tilted annuity formula Annuity Purchase_Price WACC Price_Trend 1 1 Price_Trend 1 WACC A tilted annuity is consistent with the use of the FCM method As with a standard annuity the tilted annuity should still result in charges that
76. criber numbers and gigabytes per subscriber per month The sheet performs two key functions translating the raw data supplied by TDC into Busy Hour Erlangs BHE or Busy Hour Gigabits per second and allowing for growth in volumes This sheet has a significant impact on the costs The larger the volume the larger the network usage and the lower the average cost for each product Note that the absolute volume of any one product has a limited impact unless this product has a high percentage of the entire network s volume For this reason a product with only one call minute per year can be modelled it has negligible impact on the result of the others BHE conversion factors and the average call length can have a significant impact I_Route_Table This worksheet provides a formal routing table for all of the products included in the core model It consists of two major sections Section 1 Routing Probabilities and Section 2 I Route Summary Table which feeds into C_Route_Table For each product up to three routing possibilities can be defined with percentage likelihoods attached to each one I_Route_Chart This is an interactive chart which highlights the different network elements and configurations used for each product based on the routing factors specified in I Route_Table I Building and Land This sheet contains assumptions and parameters for buildings and land and calculations of annual unit costs of land and buildings
77. ctical modelling reasons a balance needs to be sought when determining the level of detail to which different types of equipment can be defined and configured within the model This view is strengthened by the fact that different equipment suppliers can adopt different concepts in the modularity of the equipment Thus if the level of detail is too great then the model will tend to only be capable of reflecting a single manufacturer or indeed a single range from a manufacturer for each type of equipment modelled Date 30 11 2009 Page 95 Diagrams illustrating the level of configuration detail currently contained within the revised model have been included in Appendix B of this report 9 7 Network demand parameters As mentioned previously there is a need to maintain clarity over a number of parameters which are applied to the product demand along with the routing assumptions in order to calculate the aggregate and relative levels of traffic flowing across the network The first issue relates to the conversion of TDM based voice traffic minutes into IP based packets and the parameters included within the model are illustrated below Figure 9 11 Network Parameters VoIP VoIP Codec information Sampling Rate samples per second Sample Size bits Payload Frame width AmS Header Usage by Other Products Voice Payload 160 Bytes Broadband IPTV Non PSTN RTP Header ss 12 Bytes UDP Header 8 Bytes IP Header 20 Bytes VLAN
78. d LRAIC Working Group which the agency has considered in a consultation response dated 14 January 2009 released on 19 January 2009 NITA has now completed the work to clarify the content and work for other services for the review of LRAIC model for 2010 The updated note is annexed to this report Appendix D Date 30 11 2009 Page 9 2 Overview of Methodology Used to Build the Hybrid Model J LRAIC is the long run average incremental cost of providing either an increment or decrement of output which should be measured on a forward looking basis Use of the LRAIC cost estimation method therefore demands a more detailed definition of the terms long run average increment and forward looking Long run is understood as a time horizon in which all inputs including the cost of equipment are allowed to vary as a consequence of market demands Average denotes that costs connected to the production of the relevant interconnection services are divided by the total traffic in order to return an estimate of the average incremental costs There are several definitions of the term increment which is why this subject is discussed in detail below When talking about forward looking costs the actual meaning depends on what is meant by forward looking and on the computational assumptions that lie behind the optimisation function within the LRAIC method used Of special importance is the scorched
79. d installation Engineer assisted Request New setup Physical rearrangement Change of trunk number Technical assistance Cancellation fee II Calculation of attenuation on the line Unproductive fault handling Migration to full shared loop Date 30 11 2009 Page 110 Bitstream access BSA New installation Unassisted New installation Engineer assisted Reactivated installation Unassisted Reactivated installation Engineer assisted Additional work for new installation for BSA without co production Conversion from BSA on line to BSA without co production Conversion from ATM BSA to eBSA Installation of Multichannel Change of speed BSA Migration to the different BSA products from BSA without telephony from raw copper full loop from shared raw copper full loop Exchange of traffic Establishment of pre selection Establishment and operation of local points of interconnect Establishment and operation of 2Mbit s interconnection capacity Supervised access Most of these services mainly consist of the operating costs and NITA has decided to estimate the costs of the services separately The calculations are based on the activities connected to the delivery of these services The description of the tasks is based on information from the LRAIC Working Group and TDC plus NITA s reading of TDC s standard offers In response to an industry consultation in 2008 and SMP decision on ma
80. d sections refer to demand at the network level for copper and fibre respectively the second section extracts information from the first section in terms of means of transport of the different services this needs to be done because xDSL services share most of the times copper pairs with PSTN and therefore when allocating costs to services this needs to be taken into account I Dimensions This spreadsheet is a reference for the user in terms of cost and service categories used throughout the model It is useful as a summary of the dimensions of the model I Access Routing Table This table lists the access services in the vertical dimension and the network services in the horizontal dimension I ASM PDP This worksheet allows certain inputs to be defined relating to the ASM PDP links Section 1 contains a table used to assess the likely degree of trench sharing between multiple ASM PDP links The reduction factor is applied to the aggregate crow fly distance of all ASM PDP links for a particular sample area Section 2 is used to select the fibre cable used to connect the Street Cabinet DSLAM back to the ASM location for ASM PDP links I_Share_Routes This sheet specifies the amount of street trench and duct shared with other networks or utilities Routes shared with the core network are calculated in the core model and fed into the access model I GIS Trench and Duct This sheet contains geographic data focused on the road netw
81. distances on the ground Date 30 11 2009 Page 67 Subsequently the model then gt gt Determines the number of copper pairs and the number of links on each connection Determines the number of NTPs cabinets and joint boxes for each sample zone Calculates and saves these results for each of the 20 sample zones in turn This can be done automatically with the aid of a macro or the user can copy and paste the values manually if preferred Calculates the weighted sum of the results for the 20 sample zones to determine the resources needed for copper access network for the entire country in terms of length by size broken down by geotype pair km NTPs and cabinet equipment For copper cable it adjusts for spares for each connection and geotype and then distributes the entire length by size amount over the different sizes of cable through the distributional assumption Adjusts the total for the number of pair kilometres now fed by fibre More details on how the model calculates each of these steps are provided below 8 4 3 2 Average distances of different parts of the loop This section shortly describes the method used to estimate the distance in the different parts of the access network gt From ASM to PDP This distance has been estimated as the difference between the ASM and the centroid of the PDP zone the elliptic function has here been used after the original area has been converted into an EHA and the distance
82. dsloven published on the NITA website 1 3 The Hybrid Model As stated in 14 section 2 of Executive Order no 930 NITA is required to develop a LRAIC price estimation method for the following products gt Exchange of traffic between telecommunications nets or services gt Lease of Unbundled Local Loop gt Co location costs from common use of buildings switching equipment etc As a consequence of 15 of Executive Order no 930 the first version of the Hybrid Model cost base was estimated as a combination of two cost analyses gt A Top Down cost analysis conducted by the suppliers of public telecommunications nets or services who are obliged to offer interconnection products at LRAIC regulated prices in this case TDC cf 15 section 3 of Executive Order no 930 gt A Bottom Up cost analysis conducted by the suppliers of telecommunications nets or services wanting access to interconnection products at LRAIC regulated prices cf 15 section 3 of Executive Order no 930 The suppliers in question have formed a formalised network known as the LRAIC Working Group in order to conduct this analysis In accordance with 15 section 7 of Executive Order no 530 NITA established criteria and minimum requirements for the two cost analyses These criteria and requirements were laid down in a Model Reference Paper prior to the development of the original Hybrid Model This reference paper consists of
83. e gt Ifa task in practice is carried out at the same time as a task related to another service but the services are not necessarily pre requisites for each other an adequate allocation of time consumption should be made In the actual model the allocations are made by applying a weighting factor This is shown in column G of the I_ Resource sheet Date 30 11 2009 Page 112 10 2 3 Raw copper and shared raw copper 10 2 3 1 General assumptions Order processing is assumed to be automated to a substantial extent Receipt of orders happens automatically and no time is spent on receiving the order via e g phone other than in exceptional cases system breakdown etc Tasks at the exchanges are assumed to be coordinated by the supplier with the tasks at the end user premises This implies that the supplier makes use of large scale operation advantages that naturally come from the delivery of different services to both their own and other customers Transport is divided into time spent on transport following the trench to the end user and the time spent on transport to the exchange etc The purpose is to avoid double counting when the same time for transport is shared between several orders This could be the installation of raw copper shared raw copper and installation of NTP when installation of NTP is included in an order Installation of raw copper shared raw copper and installation of NTPs are independent of each other in the sen
84. e 30 11 2009 Page 80 Figure 9 1 Basic IP Network Design Core Routers Edge Routers Aggregation Switches DSLAMs MSANs The existing IP network deployed by TDC is a slightly modified version of the above in that the Core routing layer comprises two sub layers Super Core routers and Distribution routers The Distribution routers collect traffic from a number of Edge router rings Where it is necessary for one Distribution router to exchange traffic with another one it does so via one or more Super Core routers Each Super Core router site actually comprises two distinct routers for resilience reasons with the routers interconnected in a square arrangement Each Distribution router connects to two of the Super Core routers This is illustrated in Figure 9 2 Date 30 11 2009 Page 81 Layer 3 ae Le Super Core Lam Layer 3 Distribution de a gt a gt a Each Distribution router site also comprises two distinct routers to provide resilience There are no direct connections between Distribution routers Instead all relevant traffic passes u
85. e representative of the whole country covering all four geotypes and covering a representative range of zones within each geotype The most important data required for these zones is as follows gt Location of zone centre grid reference gt Location of switch grid reference gt Area in km gt Weighting i e the number of similar zones in the country Different weightings can be given to each zone in the sample gt Number of copper pairs gt Parameters that define the dispersion of subscribers within the zone gt Street level attributes the typical dimensions of a customer site such as its frontage the distance from the street duct to the customer building also whether customer sites tend to be on one side of the road or on both sides gt Finally the strategy to be adopted in designing the access network for the zone This consists of four factors two that control the number of primary and secondary distribution points PDPs and SDPs and two to control the strategy for locating them near to the subscriber near to the ASM or somewhere in between Most of the calculations for the copper network are developed for each of the 20 zones individually The purpose of these calculations is to determine the length of the local loop and then to equip it with cable In order to estimate these distances an assumption is made about how dispersed subscribers are within each ASM zone The main assumption is that each
86. e and for the country as a whole The spreadsheet also shows assumptions on unit costs asset lives and price trends 14 1 3 Output sheets O_to_Consolidation This spreadsheet lists all the results in terms of gross replacement costs split into equipment and installation costs and operating costs The spreadsheet also shows assumptions on asset lives and price trends and the network elements to which the different cost components are then allocated to In addition to national outputs results are also shown separately for each geotype O_Key Statistics This sheet summarises key inputs and calculations from different sections of the access model for the number of trench kilometres copper and fibre pairs and NTPs It then goes on to calculate statistics such as trench kilometres per NTP and copper fibre kilometres per NTP 14 2 Description of worksheets in the Core Model 14 2 1 Utility sheets In common with other models the core model contains a number of utility sheets such as Changes Version History Map and Colour Codes In addition it includes an Overview sheet which extracts key results and feeds them into the Consolidation model The Notes sheet contains a full list of modelled products and defines the names of product group prefixes and range names which will be used throughout the model Named ranges not only improve transparency within formulae but also enable users to modify individual product names in J_Product_Demand witho
87. e model are open otherwise links should be updated if other models have been changed and are located in the same directory There is an introductory message that warns the user that older versions of Excel may not provide full functionality 8 7 4 Understanding the model The recommended approach to gaining an understanding of the model is to work through each sheet in turn starting with inputs moving to calculations and then outputs One way to improve understanding is to edit the input data recalculate and investigate the effect of the changes on the results These can be easily checked in the sensitivity analysis sheet in the Consolidation model The Access model is considerably larger and hence more complicated to navigate and understand than the other models Therefore a large part of the user guide has been included in the model itself This guidance has not been duplicated in this document The final costs for all modelled access products are shown in the output and results sheets of the Consolidation model A sheet by sheet description of the Access model is provided in Appendix C of this document Date 30 11 2009 Page 78 This section describes the methodology and assumptions behind the core model This includes a description of the modelled network dimensioning of the network and the approach taken towards allocation of costs The analysis has been based on the scorced node approach as described in Section 2 2 1 above
88. e modelled Primary Distribution Points PDPs The number of PDPs in the modelled network was left unchanged and the resultant costs were based on a national average Date 30 11 2009 Page 71 In 2006 it was anticipated that TDC would be rolling out in the region of 1500 such cabinets over the course of around a two year period The reality has been that TDC has only to date introduced a very limited number of such cabinets although they do still intend to implement a more aggressive rollout programme over the coming years On the basis of past events and probable future rollouts the modelling in the revised Hybrid Model has been adapted in two specific ways gt Incorporate a facility to target specified geotype s for the Active Street Cabinets gt Adapt the assumption on PDP numbers for the sample areas such that where active street cabinets are to be deployed the number of customers fed from a PDP is around 500 at least in the more urban geotypes The model also now has the ability to include specific node equipment for the Active Street Cabinets This is on the assumption that the existing passive PDP would still exist with an Active Street Cabinet located adjacent to it The costs of the active cabinets have been based on information provided by the operators 8 4 4 2 Backhaul from Active Street Cabinet sites Specific backhaul products are now included in the revised model to cater for a number of options for active st
89. e network under the scorched node principle However where DSLAMs or MSANs have been moved from the local exchange to street cabinets under a next generation access scenario these are included in the access network classified as Active Street Cabinets After calculating the resources needed to build this network the model uses the results to calculate the capital equipment cost of those resources and the annual operating cost of such a network The model also removes from the main calculation flow the costs that are attributable to non relevant services and calculates the costs of some elements of the access network MDF line cards and other subscriber related costs The core model does not calculate the actual product costs Instead outputs in the core model are linked to the consolidation model where these calculations are carried out 9 9 2 Model structure For ease of navigation the core model includes a map of all major information flows between them This is reproduced in the diagram below Date 30 11 2009 Page 101 Figure 9 13 Core Model Navigation Map Input Calculations Output _Trenching Local _Node_ _Trenching Equipment Transit C_Trenching Local _Cost O to Consolidation _ Technical Input _Product_ Demand C_Route Table Network Elements _Route_Table ee The user can click on any of the buttons on the map to jump straight to the relevant worksh
90. e some inputs in the calculations sheets The separation of sheet types is therefore only a guiding principle However the overlaps between sheet functionality have been kept to a minimum This divisional approach to modelling promotes transparency in the calculations since any inputs are shown separately avoiding any hard coding of data within formulae Each model has a number of standard management sheets gt Workbook map or control sheet This is the normal starting point for users Workbook maps contain a diagram showing the main sheets within the workbook and the main flows of information between the sheets Clicking on any of the boxes in the diagram enables the user to move to the relevant sheet This allows easy model navigation gt Version history This records the model versions and any changes carried out Users are strongly recommended to use this and to introduce a version control procedure so that different versions are controlled gt Colour codes This contains the Excel styles and colour coding used throughout the model It can also be used for descriptive information gt Changes This provides a summary of what changes have occurred between the various versions published by NITA In addition cells in column A of individual worksheets may be colour coded to highlight recent changes gt All sheets have a macro link labelled Model Overview to link back to the model map or control sheet This eases navigation T
91. ead and allocation of cost to co location services or access as appropriate is carried out in the Consolidation model 14 4 Description of worksheets in the Consolidation Model 14 4 1 Utility sheets Control sheet Instead of model maps the Consolidation model opens in a Control sheet which is available in two language versions Danish and English This sheet contains buttons for a range of functions to aid navigation take the user directly to important inputs such as parameters overheads or cost of capital update the sensitivity analysis or print out results It also allows the user to open all linked sheets ie Access model Core model and Co location model without having to open files individually in the model directory Navigation buttons take the user to the overview maps in the other models when these are open The control sheet only functions properly if certain norms have been observed when modifying file names The model version number can be changed without consequence but more extensive name changes eg changing Draft Revised Model to Final Model may prevent the macros in the control sheet from working correctly unless the broad model version is specified in the Z Parameters sheet Section 5 To alert the user to potential problems the filename status is shown at the bottom of the control sheet Errors can be rectified either by updating the model version parameter or by modifying the control sheet macros in VBA i
92. een based on a qualitative assessment of the estimates and documentation available Where NITA has found the documentation for certain estimates to be less substantial only limited weight has been put on the estimate in question Estimates from the Swedish LRIC Hybrid Model have been used as a documented benchmark for equipment prices on several occasions This model is comprised of both publicly available data and confidential data Where the equipment price used as a benchmark is publicly available it is used without further adjustments The Swedish telecommunications regulator PTS has informed NITA that the confidential data used in the model are the actual equipment prices masked by randomly applying an uplift in the range of 10 percent Furthermore NITA has access to various international benchmark prices that may be appropiate for use in the Danish LRAIC Hybrid Model Documented and evaluated Date 30 11 2009 Page 19 price information from the Swedish model as well as other benchmarks have therefore been used as a basis of comparison with the estimates from other sources NITA has not been able to identify installation costs separately in all the equipment prices available Therefore it has been necessary to integrate installation costs in the model in one of the following three ways gt As a documented cost when information on specific installation costs has been available gt As an estimated mark up on equipment
93. eet Each subsidiary worksheet has a Model Overview button which the user can click on to return to the navigation map Most sheets also have a drop down menu below the Model Overview button to take the user to different tables contained in each worksheet Note that it is still possible to navigate the model using the sheet tabs 9 9 3 Start up There are two ways to open the Core model The first method which is recommended is to first open the Consolidation model and then select the Open Linked Sheets option which will open all the individual worksheets of the Hybrid Model This ensures that linked information is updated automatically and links are preserved when the model is saved under a different filename The user should also be aware that the Excel auditing tools will trace precedent and dependent cells only for workbooks that are currently open Alternatively the user may open the Core model directly from the file directory Enable macros when asked unless the file is from an uncertain untrusted source There may be a message asking whether to update links Such updates are performed automatically on recalculation if other parts of the model are open otherwise links should be updated if other models have been changed and are located in the same directory There is an introductory message that warns the user that older versions of Excel may not provide full functionality 9 9 4 Understanding the model The recommended approa
94. el does so via the J Interface worksheet 7 3 1 Important actions We strongly advise that all four model files should be kept in a single folder and that other files within that folder be kept to a minimum The user should track any changes to the model and utilise version control as appropriate As a minimum it is recommended that the user keep a master copy of the original models so they are always available in case changes are made that cannot be corrected or if the files are accidentally deleted The user should make full use of the facilities within Excel for exploring the workings of the models In particular the Audit toolbar is helpful in tracking the logical flow of data and dependencies The user should avoid using drag and drop to move input data since this corrupts the flow of data in the models Other than that changing the input values has no damaging effect on the workings of the model However changes to the data can make the results misleading or incorrect Entering inappropriate values e g text in place of numbers can also cause errors in the calculations giving VALUE and DIV 0 errors Some traps have been built into formulae to cope with such events but it is not practical to cover all eventualities Altering formulae is more serious and should only be done with care and after due study We would also strongly recommend making a full backup of the models prior to changing the formulae just in case Partic
95. ement System Core infrastructure management and planning Core infrastructure maintenance Core DWDM equipment management and planning Core DWDM equipment maintenance IP network management and planning IP network maintenance DSLAM management and planning DSLAM maintenance Access infrastrucuture management and planning Access infrastructure maintenance IMS installation Core infrastructure installation Core DWDM installation IP Network installation DSLAM installation Access infrastructure installation Corporate Overheads Human resources Finance Support systems Administration Customer oriented costs Billing Debtor handling Other IC specific costs The cost categories summarised in Table 3 2 represent a significant departure from those used in version 2 4 of the Hybrid Model This was necessary to take account of the considerable differences in the underlying network structure and hierarchy resulting from the adoption of an IP based core network Costs are split into continuous and one time costs Personnel demand is calculated per FA area and the total wage costs are calculated using a standard price for the pre defined position categories This is based on the following three standardised personnel categories gt Technicians gt Administration e g secretaries gt Academics professionals such as solicitors engineers etc NITA has estimated an average yearly gross w
96. en changed and are located in the same directory There is an introductory message that warns the user that older versions of Excel may not provide full functionality 10 3 4 Understanding the model The recommended approach to gaining an understanding of the model is to work through each sheet in turn starting with inputs moving to calculations and then outputs One way to improve understanding is to edit the input data recalculate and investigate the effect of the changes on the results These can be easily checked in the sensitivity analysis sheet in the Consolidation model The final costs for all modelled co location and other products are shown in the output and results sheets of the Consolidation model A sheet by sheet description of the Co location model is provided in Appendix C of this document Date 30 11 2009 Page 120 This section sets out to describe the methodology and assumptions behind the consolidation model and describes the approach taken to the modelling of operating costs and also the production of the final cost based prices 11 1 Overview of the Consolidation Model The consolidation model uses the costs relating to core access and co location that are produced by each of the separate models The costs are brought together in the consolidation model The consolidation model produces the final cost of each service It also undertakes some checks for consistency between the other three models In contrast to prev
97. en rounded to the nearest number divisable by 5 This is however not the case for asset lives that are shorter than 10 years Date 30 11 2009 Page 38 Table 4 2 Asset lives Cost category Hybrid Model Access duct 40 Access trench 40 Copper cable 20 Access fibre 20 Cabinets distribution points 15 NTPs 10 Line cards 5 Circuit boards jointing 15 DSLAMs MSANs 5 Layer 2 processors 5 Layer 3 processors 5 Peering routers 5 Media gateways 5 IMS controllers servers 5 IN platform 5 MDF 15 International switch 10 DWDM transmission 5 Power supply unit backup 15 Air conditioning unit 15 Security system 10 Security guard 0 Site preparation 10 Site maintenance 0 Core duct 40 Core trench 40 Core cable 20 Submarine cable 15 Technical house rack space 10 POI 10 Buildings land 30 IT cabling and PCs 6 Asset lives for trench and duct in soil is 20 years as the cable is ploughed in the ground Date 30 11 2009 Page 39 4 6 Price trends Price trends used in the model should be forward looking indicating the expected future development in equipment prices as the trends have to be used to project the equipment prices in the model Using the historic development in prices as a proxy may often be the best way to estimate future price changes Such a methodology can only be used however if the past development is believed to co
98. ent at a certain site the SMP operator is obliged to provide the necessary transmission capacity free of charge from the Operator s premises and to the building of the co location 10 1 2 Modelling Co location The primary purpose of the co location model is to cost existing co location services used for access to the unbundled loop Therefore the model considers costs that would be borne by the SMP operator in the event of offering co location services at appropriate sites in the network Unlike services in the core and access network co location services consist of relatively few cost categories These can be divided into costs that are specific to a particular service and into costs that are shared with other services 7 Examples of demarcations include solid walls lockable cabinets or wire meshes Date 30 11 2009 Page 105 Services in the core and access network are costed by combining costs from a pool of cost categories using a routing style allocation table technique This is not the case with co location products These are mostly standalone sub products that may be combined by the operator who demands co location Therefore although the co location model is simpler in structure compared to both the core and access models special care must be taken with the co location model in order to model and capture costs at a sufficiently granular level One of the most important differences between the co location model and the a
99. er and shared raw copper and transport e g when a wire has to be moved from a multiplexer to raw copper or be pieced together via several stretches gt Time for transport covers the transport to and from the exchange but because new setup is a service that is carried out in connection with the installation the transport is already covered by this service Transportation is widely covered by the installation Additionally it is understood that the technicians are based locally which also reduces the need for transport Therefore a correction for double counting of time for transport the share covered by new setup is made in relation to the services installation of raw copper shared raw copper and other conditions reducing the need for transportation 10 2 3 5 Physical rearrangement Physical rearrangement of raw copper is regarded as a service that is carried out as an extension of installation of raw copper In that light NITA has found the following cost driving activities gt Processing of order is estimated to be covered by the ordering of the raw copper or shared raw copper and thus it only contains the extra time spent in connection with the physical rearrangement gt Physical coupling in exchange covers the actual physical rearrangement from one pair to another connected to the installation of raw copper Date 30 11 2009 Page 114 gt Time for transport covers the transport to and from the exchange but as
100. ersion of the LRAIC Hybrid Model For data inputs other than unit costs NITA has whenever possible followed the same principles as when dealing with unit costs However in one instance it has not been possible to let the actual data inputs be accessible to all parties since they are highly dependent on confidential data supplied by TDC As a consequence NITA has chosen to submit a public version of the model containing masked data instead of the underlying values used in the model This is done by adjusting the actual data inputs to the model by a random factor negative or positive In the published version of the Hybrid Model it is indicated explicitly by the use of a specific colour code when data is masked Date 30 11 2009 Page 20 The very limited use of masked data makes it possible to use the model with almost accurate data input The random adjustments to the published model produce only very small deviations from the model producing the correct non masked data 3 4 Operating costs In previous publications NITA has acknowledged the difficulty in modelling operating costs in a bottom up model and noted that only sparse documentation and justification had been provided for the mark ups used in earlier versions of the model To improve transparency and model operating expenditures more directly NITA chose to adopt a Functional Area approach in the Hybrid Model which was based on the FA method used in the Swedish LR
101. ervice a distinction is made between the time spent by administrative personnel academic professional personnel and technicians To calculate the cost of delivering the service the estimated time spent is multiplied by the hourly rate outlined in the LRAIC model The hourly rates for administrative personnel academic personnel and technicians are shown in Chapter 3 For co location NITA uses the same hourly rate as in the access model As for other co location services NITA has applied a mark up to ancillary products to take account of relevant overheads This is added in the consolidation model 10 2 2 Handling double counting of order processing and transport In relation to each service NITA has evaluated to which extent a task is a natural extension of another service and to which extent transport is shared with other services Moreover it has evaluated whether the SMP operator with reason might bundle orders within the same geographical area and thus share the time for transport between several orders The following principles have been applied gt Ifa task in relation to service B is always carried out as an extension of a task related to service A and service A is a technical pre requisite for service B then service B should only contain the incremental cost derived from the extra work process undertaken in relation to service B This implies that no party can be brought to pay for tasks not related to the ordered servic
102. et is considered to be around 500 Fibre The modelled amount of fibre used to connect business customers has been derived by modifying the existing fibre requirements using MEA adjustments Additional fibre is included to connect the exchange locations to the acive street cabinets Subscribers connected through copper are assumed to continue to be supplied by copper unless the NGA Selector is switched on in which case the model provides an illustration of the costs for a full fibre network NTPs Network Termination Points have been modelled on the basis of the existing number of lines served in the network Line cards Line cards have been modelled on the basis of the existing number of lines served in the network Date 30 11 2009 Page 60 8 3 Main definitions and assumptions in Access The Main Distribution Framework MDF areas in the access model are allocated to one of four categories called geotypes according to their tele density gt City Storby more than 1000 lines per km2 gt Urban By 100 to 1000 lines per km2 gt Rural A Land A 10 to 100 lines per km2 gt Rural B Land B less than 10 lines per km2 The optimal design of the network and the mix of costs incurred will be quite different in each of these types of area It will also vary although to a lesser degree within a geotype since two zones of the same geotype can still have very different tele dens
103. eters built into the model that the various industry participants can relate to and therefore comment on during the various consultation processes 9 8 Detailed description of cost calculations 9 8 1 Overview Gross Replacement Costs for exchange equipment are calculated by multiplying the equipment prices in the sheet I Costs with the amount of equipment required This takes place in worksheet C_Equipment for the active equipment and C_Trenching Transit and C_Trenching Local for the trench duct and fibre infrastructure Some of these costs are classified as call related while others are volume or subscriber related The necessary transmission capacity in the network covers the demand for voice and data services leased lines and other services such as IPTV The actual demand for Date 30 11 2009 Page 97 each product or service is fed from I Product Demand into worksheet C_ Route Table which converts the demand for products into demand for network elements required to provide these products To allocate costs to core services the model gives users the option to select from five different units of usage gt busy hour calls BHC gt busy hour minutes measured in Erlangs BHE gt busy hour Gigabits per second BH Gbps gt busy hour Mega packets per second BH Mpps gt number of subscribers Subs Using busy hour minutes as the allocation key the product s average use of a network element is divided by t
104. ethods a reasonability and or consistency check is performed by e g comparing with the numbers from the current LRAIC model to see whether a correction should be made The relationship between personnel and non personnel costs is predominantly based on TDC s data from the original top down model It should be noted that the network based FA costs do not include annual costs for premises power and cooling These costs are modelled directly in the core model and then transferred to the consolidation model Date 30 11 2009 Page 27 3 4 3 Working capital The cost of the working capital is a percentage of the total working capital The percentage value used is assumed to be the same Cost of Capital percentage that is used to define the return on fixed assets The same modelling method is used as was in the previous version The Required Level of Working Capital RLWC is defined as the current assets less current liabilities Thus 1 RLWC Stock debtors creditors cash Stock is assumed to be negligible 1 can subsequently be calculated as 2 RLWC Debtor days sales creditor days total creditor related costs 365 cash Sales are the sum of the sales revenue of Co location Core PSTN and Access network services These are calculated in the model assuming a cost oriented sales price is used The total trade creditor related costs are made up of gt Wages gt Power gt Other payments to suppliers
105. evels At the lowest level one or more DSLAMs are connected to Layer 2 Ethernet switches These switches are capable of the vlan tagging necessary a to allow different data streams to have different priorities over the network and b to allow multiple OAOs to gather traffic from their own end users connected to the DSLAMs covered by that switch At the next level a number of Layer 2 Ethernet switches in a geographic area are connected in a ring to a Layer 3 IP router If an OAO connects at the location of this Layer 3 router then it would have access to all DSLAMs covered by that router The routers are then connected together in higher level regional rings to IP MPLS core routers The highest level of the network consists of the interconnection of these MPLS routers OAOs are of the opinion that modern Ethernet based DSLAMs could support direct connections from multiple operators TDC has accepted that this might be the case in theory although they contend that the maximum number would be limited but believe that for overall operational reasons interconnection should be achieved via a Layer 2 switch Date 30 11 2009 Page 47 Figure 5 3 Functional block overview of Ethernet based xDSL network Layer 2 nodes in ring structures
106. f subscribers and the volume of traffic will be correlated Nevertheless it is possible to consider the implications of increased volume of traffic by keeping the number of subscribers constant implying an increase in the usage rate or to consider an increased number of subscribers by keeping the volume of traffic constant implying a decrease in the usage rate Assets within the core network typically include gt Exchanges and exchange equipment e g traffic related costs of DSLAMs MSANs core and edge routers aggregration switches media gateways gt Transmission links between the exchanges IP network DWDM gt Trenching and duct between all levels of exchanges 2 1 2 2 Access As defined above costs in the access network primarily depend on the number of subscribers and only to a very limited extent on the number of calls or volume of traffic Consistent with this an alternative definition of access is that it is the service allowing the subscriber to send and receive traffic Both definitions suggest that the access network includes all cable and trenching costs associated with subscriber lines between the customer s premises and the concentrator DSLAM The definitions also suggest that the access network includes analogous costs for other lines such as those for public call boxes between the customer s premises and the concentrator Furthermore the definitions suggest that the access network includes subscriber related
107. f the 20 sample areas and over the whole network using the DAV database The numbers from the DAV database are also corrected using information from Statistics Denmark about the number of inhabited residences gt by conducting a stand alone analysis of the average distance between sites and road side by geotype for a sample of streets Raster maps as indicated in the following table were requested from TDC to assist the calculation of the distance from the exit to street duct EFSD to customer premises In order to ensure a reasonable degree of accuracy in the sample six areas were selected within geotype Storby 10 within geotype By 18 within geotype Land A and six within geotype Land B These numbers were chosen to reflect the relative diversity of geotypes in Denmark Once the different zones had been selected streets were chosen randomly within these zones For the selected streets in each geotype 10 house numbers were randomly chosen and two distances measured gt The shortest distance beeline from the street duct to the house gt the actual trench distance from the street duct to the house Under this approach the number of houses sharing the trench was noted and the calculations adjusted to reflect this sharing If for example the exit from street duct was shared by two houses the total trench length was measured and divided by 2 The two approaches led to similar results the former approach of
108. f the conversion factors through an analysis of TDC INCA maps Through these maps NITA looked at the amount of trench in place in TDC s network for a sample of roads that would try to mirror the different road classes previously identified i e AA AB etc In many cases the actual network was used to set the conversion factor However in some areas particularly those with very few sites an adjustment was made to reduce the amount of trench in the network where trench was considered to be excessive for the area served Moreover when the analysis was carried out it was not always possible to extract from the available maps a meaningful sample of roads for each of the 10 classes of Date 30 11 2009 Page 62 road Regarding those classes for which it has been considered that the available information was not sufficient to estimate the conversion factors an econometric analysis has been applied instead This was undertaken by estimating the remaining conversion factors and the associated weights for the A B C and D segments The conversion factor for the CD CD segment consists thus of the weighted average of A and B with the value of 0 and C and D with the value 1 The estimated conversion factors are thus adjusted to the directly measured conversion factors Concerns were also raised during the original reconciliation phase with respect to the treatment of so called Ekstra vej extra roads The conversion factor associ
109. h other gt 108 edge node sites 13 shared with the distribution node sites These are connected together in logical rings with each ring having two distribution node site heads and a total of between 5 and 7 sites on each logical ring Accepting that a logical ring might pass through physical locations not included on that ring the model allows the user to specify where such locations are used with the fibre assumed to pass through that location but not requiring any active equipment at the location gt 1 838 DSLAM MSAN node sites 108 shared with the edge node sites These are again connected together in logical rings although due to the number of sites involved this is not carried out ring by ring within the model Instead a design maximum number of Layer 2 sites on each logical ring is specified in the model and used to calculate the required number of such rings to connect each edge node site with its relevant DSLAM MSAN node sites Once this overlay is complete it is then possible for the model to overlay the traffic onto these logical rings which in turn allows the model to determine fibre counts and necessary IP equipment configurations The model contains a complete set of physical route segments and all logical route segments allowing modification of the overall design according to differing principles should this prove necessary in the future 9 6 Modelling the core network 9 6 1 Voice telephony overlay TDC
110. he LRAIC Hybrid Model is based on a Bottom up Cost Model which has been built according to a number of specifications or assumptions The most important of these is the scorched node assumption An optimal network structure has thus been built under the restriction that all exchange sites in the existing network are populated with equipment however this will be modified as a consequence of more effective technology The use of the scorched node assumption is described below 2 2 1 The scorched node assumption The legislation in Denmark requires the LRAIC price estimation method to adopt the scorched node approach to modelling In 2 section 4 of the Executive Order it is stated that The cost analyses shall start from the basic geographic network structure at providers of public electronic communications networks or services that are subject to price regulation following the LRAIC price estimation method This implies that exchange equipment or similar equipment shall be placed at the geographical sites of the providers network architecture scorched node This implies that the optimisation in the LRAIC Hybrid Model is constrained by the existing number of sites and their geographical locations The scorched node assumption does not imply that the transit net cable duct trench etc is constant Nor does it imply that the geographical sites shall be populated with the same number or types of exchanges as in the e
111. he normal Excel navigation tabs can still be used A drop down menu takes the user to individual tables within each worksheet The models all have the same Excel Styles and colour coding This ensures that the cells are easily identified The Styles used are shown in the diagram below Date 30 11 2009 Page 53 Can the user edit it Meanina Looks like Yes Options that will affect results Bold Green text pale green background Yes Data that generally affects results Yes Input data that is a results of off line Bold Green text pale green blue background analysis or data where care must be taken when making changes L Yes Comments about the data No Green text lilac background direct impact on results Shows the structure of the Bold Blue text pale grey background calculations Sub header of table Shows the structure of the Bold Black text clear background calculations Header of sub table Shows the structure of the Italic black text white background calculations Calculations Calculations that generally affect Black text clear background results Developer notes Design build issues Bold Purple text clear background no just take note Warnings and errors Error gt gt gt Bold red text clear background means it needs fixing Cells that link to ie Calculation cells that refer to other luse External data workbooks Camouflaged for confidentiality Camouflaged data reasons Bold White text bl
112. he number of busy days per week a user input This is further divided by the number of busy hours per day also a user input In the previous model these parameters were hard wired into the conversion algorithm The introduction of user inputs allows the model to adapt to evolving trends such as increasing broadband usage per day or the fact that traffic loads are no longer determined predominantly by office hours or peak off peak tariffs Finally a factor of 60 is applied to convert the number of minutes to hours The parameter default settings are guided by information provided by TDC 9 4 Key Network Elements The list of Network Elements used within the core model has been by necessity completely revised to adapt to the new IP network structure In arriving at the definitions of the Network Elements a balance was sought between the need to gt Keep the number of Network Elements to a manageable total gt Allow for allocation of costs in as cost causal a manner as practical gt Maintain a single unit of use for each Network Element The selection of Network Elements also had to reflect the level of available detail in the modelled equipment configurations there being little point in having Network Elements at a greater level of detail than that in which the equipment could be usefully Date 30 11 2009 Page 90 defined within the model The revised list of available Network Elements together with their respective units of
113. he total volume in minutes through the element For busy hour Gigabits per second and busy hour Mega packets per second the model uses each product s share of total Gigabits per second or Mega packets per second traffic in the busy hour as the allocation key and so on As described in Section 9 7 above the model uses conversion factors which allow Mega packets per second to be calculated from Gigabits per second with the exception of VoIP traffic where the packets per second is an input parameter dependent on the selected Codec The amount of equipment required is further determined by the number of nodes in the TDC network based on the scorched node assumption and the distances between them The distances between nodes have been determined by the Autoroute trenching analysis as described in Section 9 5 above The results are captured in the trenching input sheets in the core model which feed into C_Trenching Transit and C_Trenching Local where the cost of trenching and cabling are computed The cost of other network elements such as routers and switches is calculated in C_Equipment The results from the equipment and transmission cost sheets are summed up in O Consolidation which feeds into the Consolidation model 9 8 2 Cost input Costs for different sizes of network elements are estimated in order to increase model accuracy flexibility and to allow creditable cost volume relationships Equipment cost inputs are broke
114. hours per day for technician i 7 4 Average number of hours per day academic staff Hours a 7 9 Correction for additional work 25 0 5 2 Average employment costs Per year Mark up Administrative staff DKK pa From consolidation m Technician DKK pa From consolidation m Academic staff DKK pa From consolidation m 4 empty DKK pa I empty DKK pa p me Source Co location model I Costs Any use of time on activities that are not directly work related such as transport waiting time and other relevant wasted time is taken into account when calculating time usage for each specific activity e g installation of raw copper Moreover new service categories have been added in the revised Hybrid Model to account for unsuccessful customer visits and unproductive fault handling Regarding services tasks in which area the estimated cost is based on the number of hours and wage level a mark up of the hourly wage is needed to take into account the indirect costs derived from other types of personnel who indirectly contribute to time expenditure for that activity For example a problem on an access line will primarily need a technician who will use a certain amount of time on repairs and testing but the same problem will also result in extra use of administrative and managerial hours Similarly an academic employee negotiating with a provider about a relevant interconnection contract will also draw upon administrative and managerial resources
115. i Full Loop see eeeseemsemneneenes Sub Loop ae In theory the interconnection point could be at any junction point in the local loop cabling However in practice it is likely that the interconnection will take place at the larger street cabinets since the smaller street cabinets are unlikely to provide sufficient numbers of subscribers to make the unbundling worthwhile to the OAO Following discussions with NITA it was agreed to consider that a sub loop would run between the PDP Primary Distribution Point and the end user s NTP In particular it Date 30 11 2009 Page 45 was also agreed that the definition of what constituted a PDP would be based on the existing LRAIC model and furthermore that every PDP within the model would be considered a potential sub loop unbundling point The latter agreement was deemed necessary to remove any reliance on TDC s own Street Cabinet DSLAM deployment programme particularly as there was no available source within TDC of which street cabinets would be upgraded and in any event these locations would most likely not match those included within the LRAIC model since the latter took a Bottom Up approach to their location Version 2 3 of the LRAIC model modelled full local loops in two distinct parts trench duct and cabling The trench and duct required was calculated on a national basis using a database of the Danish road system The cabling requirements were calcul
116. individual repairs and common goods such as development of the network gt Furthermore a non pay operating cost is modelled for the necessary expansion of the NMC system This is calculated as a direct annual costs equal to the licence fee for the NMC system 10 2 7 2 Interconnection capacity Interconnection capacity is a 2 Mbit s port 30 group in the co location exchange Installation The cost drivers for installation of interconnection capacity are gt Processing of order includes tasks for administrative staff e g processing of order confirmation registration of order and ordering hardware gt Mounting of hardware also including the physical coupling done by a technician Annual Cost The annual cost of interconnection capacity consists of two types of costs gt An annualised cost of ETC hardware trunk port gt Maintenance costs The price of ETC hardware is included in the model via a link to the core model where the price of trunk cards is included as an input The annual cost of ETC hardware is annualised based on an asset life of 10 years with the same price trend as for ports The maintenance costs are linked to calculations in the core mode where they are calculated with the annual cost 10 2 8 Supervised access with notice Supervised access with the participation of a technician is charged by the hour using the hourly rate for technicians including overhead from the LRAIC model
117. ing happen that requires operating costs to be incurred gt How much operating cost does each event cause Date 30 11 2009 Page 136 These are multiplied together then multiplied by the volume driver to derive the cost per year The results feed through to C_Resources 14 1 2 Calculation sheets C_Trench by Zone This is worksheet performs the calculations necessary to split the trench and duct costs between ASM PDP links and PDP NTP links Section Sample Zones contains road lengths by segment classification for each of the 20 sample areas In Section 2 Trench km by road type this information is then used together with the relevant road to trench conversion factors to calculate the normalised trench length for each weighted sample area The total trench length estimated for each weighted sample area needs then to be allocated between the primary ASM PDP and the secondary network PDP NTP This has been done by calculating the estimated length of the primary network and subtracting this estimate from the total estimate in order to obtain the estimated length of the secondary network Section 4 Mix of Duct types is an expanded equivalent to Section 4 of worksheet I GIS Trench and Duct and Section 4 of worksheet C_Trench and Duct The note detailing how the amount of Large Duct was estimated has been used to calculate values for each of the 20 sample areas These values are normalised to give the 500 km inputted in worksheet J_
118. ing the value of the land property area in m land area in m and postcode for each site Information concerning the cost of land is primarily based on the public assessment from 2004 however in a few cases the details for 2004 were not available here the valuation from 2002 has been used In order to correct the public assessment property data from TDC NITA obtained data from Statistics Denmark illustrating the difference between market prices for factory and storage properties and the public valuation through four quarters from the second quarter in 2004 through to the first quarter in 2005 For version 2 5 of the model updated market values were obtained from Denmark Statistics for the period from the second quarter 2005 to the second quarter 2008 the most recent available to bring the TDC data up to date NITA has calculated a correction factor at county level whereby it has been possible to make an adjustment for the public assessment value data from TDC for exchanges and technical houses in the counties with sufficient data available Building costs per m are calculated in the hybrid model and have simultaneously been split into indoor and outdoor costs primarily because land and buildings have Date 30 11 2009 Page 29 different salvage values asset lives and price trends Another reason is so that the outdoor co location area can be modelled The following calculations have been performed gt The average buildi
119. ions over long distances gt To facilitate the provision of very high bandwidth circuits that have no intrinsic need to utilise equipment within the IP network Date 30 11 2009 Page 94 gt To provide resilient fibre routings in the event of cable failure Whilst no specific information was available regarding the second of the above bullets there is still merit in including DWDM equipment within the revised model in order to address the first bullet This would in turn better reflect the type of network architecture currently being deployed The current distance limit for fibre modules in Layer 2 switches and Layer 3 routers tends to be around 80km beyond which there is a need to repeat or regenerate the signal DWDM equipment typically provides modules that can communicate over greater distances 100km unamplified is assumed in the revised model but more importantly allows for simple and cheaper amplification of all wavelengths carried by a fibre whenever that is necessary without having to turn the optical signal back into an electronic electrical one first as would be the case with a conventional repeater regenerator An analysis of all links greater than the 80km limit was conducted and a conceptual network of DWDM links that would cover those routes developed This is illustrated in the diagram below which is also included as a worksheet within the core model Figure 9 10 DWDM Link Structure This graphic ha
120. ions will be re sized in accordance with the Active Street Cabinet assumptions that is a design guideline of around 500 customers per PDP particularly in urban locations gt Active Street Cabinets will be introduced at all PDP locations and will be fed with high fibre count cable though not sufficient fibres for a full fibre to the home scenario from the existing DSLAM MSAN sites gt A full fibre to the home rollout will be assumed from the Active Street Cabinet This is in contrast to assuming a Passive Optic Network PON architecture The rationale for assuming a non PON architecture from the PDP site is twofold gt From a bottom up perspective it will allow a greater degree of future proofing than a PON architecture gt It will facilitate fibre loop unbundling albeit at the street cabinet level Date 30 11 2009 Page 75 8 7 How to use the Access Model 8 7 1 Model scope The purpose of the access model is to calculate the resources required to build that part of the network that is below the existing MDF site housing also a Local Exchange DSLAM or MSAN as far as the Network Termination Point NTP at the customer premises The model does not calculate access network costs at the exchange site itself as the following diagram illustrates Figure 8 4 Scope of Access Model MDF site Line cards Access Network Model Customer Site NTP Any parts
121. ious versions the version 2 5 of the revised Hybrid Model shows geographically disaggregated costs of access products and services through to the Output sheet Costs are now allocated to products for each geotype individually and these are then aggregated to calculate the total national costs The consolidation model performs a number of additional functions such as annualisation of capital expenditure and the calculation of Functional Area costs These have been described in detail in Chapters 3 and 4 of this report As an alternative to annualisation any investment cost may be expensed where the cost is recovered as a one off payment Some cost items that are relevant to access or co location services are typically treated in this way Core PSTN or IP network costs would not normally be expensed The model therefore calculates service costs as a mixture of one off and annualised costs Finally the consolidation model includes a sensitivity analysis which allows users to test the sensitivity of any model changes on final outputs Further details of the functions performed by the Consolidation model are provided in the following sections 11 2 Operating costs The Functional Area approach The overall rationale behind the functional area approach was dealt with in Section 3 4 Thus this section concentrates on describing the actual calculation steps themselves There are two specific worksheets that deal with the functional area costs
122. is then estimated to be circular A rectangular and a circular SDP area can be identified in the above listed figure The two sanity checks are gt The length and width of the SDP zone is compared with the diameter of the EHA for the PDP in question The length and width of the SDP zone should be smaller than the diameter of the EHA for the PDP in question gt The average distance between the subscriber and the SDP cage is compared with the average distance between the subscriber and the PDP cage The average distance between the subscriber and the SDP cage should be smaller than the average distance between the subscriber and the PDP cage If neither of these conditions is met the SDP zone is determined to be circular rather than rectangular Date 30 11 2009 Page 69 8 4 3 3 Beeline adjustment factor The beelines are converted to actual distances by multiplying each distance by an adjustment factor These factors are model inputs and vary according to the different parts of the network and the geotype 8 4 4 Street cabinets NITA decided that the average number of lines per SDP to be assumed should be in accordance with the experience from other EU countries NITA has thereby obtained access to information on the average number of PDP and SDP cabinets per MDF area and in the total network as well In addition information has also been received regarding the average number of subscriber lines in use per PDP and SDP NITA als
123. isting LRAIC model to facilitate the determination of interconnection prices for bitstream access BSA which is the description used within Denmark for the resale of xDSL connection via TDC s network Certain aspects of the equipment and physical telecommunications network required for the production of BSA had already been modelled in the previous LRAIC model This was primarily the case for raw copper and exchange buildings However certain equipment related specifically to the production of BSA needed to be added in Following discussions with NITA it was initially agreed that three separate BSA products would be specifically modelled BSA at the Street Cabinet BSA at the Layer 2 Ethernet switch location and BSA at the Layer 3 IP router location It was not required for the model to include a cost estimate for BSA National Access or indeed retail xDSL services except in so far as it would be necessary to include them for the purposes of sharing of common network elements such as DSLAMs In addition to expanding the model to include BSA products NITA required that the LRAIC model should also be upgraded to cover sub loops of raw copper Following discussions with NITA and the parties at the outset of the project it was agreed that a sub loop would be defined as the copper pair connection between the Primary Distribution Point PDP and the Network Termination Point NTP 5 2 Impact on access and core models In version 2 3 of the model
124. ithin the core and access increments The first method is to look at costs that are directly attributable to services within the increment For example ADSL line cards are directly attributable to the xDSL part of the access service within the access increment However because of the significant extent of fixed and shared costs in telecommunications networks there will be relatively few directly attributable costs Most costs will be considered to be shared costs that is shared by a number of different services in the increment such as leased lines and voice traffic Common costs arise only when two increments for example access and core actually share costs This is discussed in Section 2 1 1 above Another method to use when looking at costs is to consider how they relate to the network For example costs could be defined as gt Direct network costs such as processors ports ducts and fibres gt Indirect network costs such as power accommodation and maintenance gt Interconnection specific costs such as billing and billing systems gt Overheads such as the human resources department Network costs measure the cost of those inputs necessary for the network to run They can be divided into direct and indirect network costs A direct network cost is defined as one where the volume of inputs and therefore the cost depends on factors exogenous to the network such as the level of demand For example the
125. ities Although the Hybrid Model currently calculates geographically averaged results the revised model also shows geographically de averaged costs for the four geotype categories This facilitates analysis of geographical de averaging in order to inform future pricing decisions The main assumptions in the access model are as follows gt In general subscribers are connected to the same ASM site as at present gt Sharing of trench and or duct is allowed with the core network with other TDC increments such as its cable TV network and with other cable TV networks and utilities gt There is no sharing of copper cable with the core network with other TDC increments or other utilities In addition the access model includes technical houses which are also found in TDC s network The purpose of these technical houses is to reduce the loop lengths for the relatively small number of subscribers connected to these houses In that sense they are similar in concept to the active street cabinets although they are not modelled as such 8 4 Modelling the access network 8 4 1 Trench and duct The network of trenches and ducts is assumed to follow the road network The primary input data is therefore a breakdown of the national road network This is done by road type and by geotype The classification by road type refers to sixteen categories of road which are defined according to the number of addresses found on each side of the
126. l I FA_Costs The initial entries in the functional area staffing input table have been based on an assessment of the staffing levels in version 2 4 together with experience of inputs used in other such models The overall level of staff covered by the complete set of functional areas has been kept broadly constant with the total in version 2 4 1 435 in the release for 2009 of version 2 4 compared with 1 426 in the final revised model version 2 5 The complete new list of functional areas together with the staffing level input drivers is illustrated in Table 3 5 Date 30 11 2009 Page 24 Table 3 5 Functional Area Staffing Levels 1 2 Areas relating to ongoing and one off organic growth costs Minimum fle per driver Ongoing IMS management persons a ooa IMS IMS maintenance persons C sof 100 IMS IMS planning persons Oooo wo wo zws Network Management System persons L 150 Core infrastructure management persons ee a a 1 km Core trench Core infrastructure maintenance persons km Core trench Core infrastructure planning persons Core DWDM equipment management persons Core DWDM equipment maintenance persons 50 100 DWDM Core DWDM equipment planning persons Coo o S IP network management persons o 100 S IP network maintenance persons oOo 20 00 1 00 0 00 Core nodes IP network planning persons Coo wo DSLAM management persons 10 00 00 DSLAMs DSLAM maintenance persons 400000 1 00
127. l so that they can be taken into account there also 9 8 3 2 Allocation of shared costs for other services The IP network is dimensioned to carry a certain amount of traffic which is not modelled explicitly such as leased lines mobile or cable TV This has been grouped together under the category of non PSTN traffic and should carry a reasonable share of the cost of using the transmission network This issue has been addressed in the core model by modelling one or more aggregated non PSTN products Date 30 11 2009 Page 99 The key to allocation of transmission costs driven by the network is based on the total traffic including a possible positive growth rate In practice this means that costs equivalent to the traffic in the costing year i e the maximum traffic because growth is negative are allocated to modelled products With regards to other transmission and infrastructure costs these are allocated on the basis of the actual traffic in the costing year Infrastructure costs are driven primarily by cable length and the number of nodes and not directly by traffic Therefore these are roughly fixed in the long run For these cost categories cost is allocated between modelled products and non PSTN services based on the actual volumes in the costing year and not volumes incl growth 9 8 3 3 Allocation of shared infrastructure costs The amount of core trench shared with the access network is calculated in order to determi
128. leased lines and gt any other services including those offered by other operators to end users via the SMP operator s network An assumed rate of growth over the assumed planning period is added to the current volume of traffic in order to attain end user demand The decision of which growth rate to use is based on an assessment of the time factor involved in building the network The planned expansion should thus take into account the expected time used for the individual activities preceding the date when new major or minor parts of the core network are put into use This particularly concerns activities such as obtaining local planning permits construction acquisition installation testing operationalisation of equipment for expansion or new installation exchanges transmission equipment infrastructure etc It should be noted that the model contains a number of negative growth rates If the network was dimensioned in order to carry only the future lower traffic demand it would be under dimensioned to carry current traffic Instead the network should be dimensioned to be capable of carrying the largest amount of traffic within the planning period Therefore a correction mechanism has been built into the model which ensures that under dimensioning does not occur By using the routing factor method see Section 9 3 2 below the network is subsequently adjusted to the dimensioned demand which the network has to car
129. licing boxes are not directly estimated in the model it is possible to estimate the impact of the first factor This has been done in the following manner gt The cable lengths for the different cable sizes have been calculated from the utilisation rates in the earlier top down model gt The distribution of cable sizes is equal to the distribution in the top down model gt The distribution is skewed slightly to the left thinner cables replacing thicker cables until the cable lengths ranked after size for each geotype corresponds to the higher utilisation rates applied in the hybrid model 8 4 5 Fibre access network For business subscribers having fibre feeds the model works from an existing set of TDC data on the number of relevant subscribers and the amount of fibre in the access network required connecting them The steps are gt Define a Calibration Scenario using historical data a known number of connections and a known quantity and mix of installed fibre gt Convert the mix of fibre into Modern Equivalent Assets MEA gt Define the number of connections in the Actual scenario gt Define the Cost volume Relationships CVR for fibre in rings and fibre on spurs gt Calculate the fibre required to fulfil the actual scenario by starting from the calibration scenario and using the cost volume relationships to adjust for the differences Additional fibre is included within the modelling to account for the need to c
130. look up technique is used widely in the model to save on processing power and reduce the size of the overall model All the outputs from this worksheet feed directly into the Consolidation model 14 3 Description of worksheets in the Co location Model 14 3 1 Input sheets I Interface This sheet contains inputs from other parts of the model that are shared with the co location model via Excel links It provides information about the original input source to allow the user to locate the yellow or blue modifiable input cell I Costs This sheet contains the various cost assumptions for land and buildings co location site costs power and cabling In Section 5 it provides inputs for assumptions which make up hourly wage costs such as working day assumptions and annual employment costs fed in from the Consolidation model for administrative staff technicians and academic professional staff Section 7 lists the service elements used in the co location model The prefix used for the full service element name indicates whether final costs will be allocated to co location eg rack space or access eg raw copper installation A small number of services such as preselection local POI and interconnection capacity are allocated to other services I Resource sheet This sheet is used to define the assumptions on time resource requirements for various co location activities Vertically the I Resource sheet is divided into different sections
131. mal working hours the hourly rate for technicians incl overhead calculated in the LRAIC model is used Generally the LRAIC calculated hourly rate for technicians is the standard offer for raw copper and shared raw copper 10 2 3 8 Cancellation fee Il According to TDC s standard offer for raw copper and shared raw copper a cancellation fee II will be paid if an operator cancels an order after receipt of order confirmation with the date of TDC customer service until no later than midday 12 00 the working day before the date of delivery 10 2 3 9 Calculation of attenuation on the line An attenuation calculation which is theoretical calculation of the wire pair attenuation on the basis of TDC s cable can be ordered either when carrying out an installation of raw copper or an already established raw copper connection 10 2 3 10 Migration to raw copper shared copper When sought a conversion from shared raw copper to administrative raw copper there is a administrative charge paid The following migration services to LLU is calculated 1 Migration to raw copper full loop from BSA visit and transport to the end customer is not necessary from shared raw copper only order processing necessary Date 30 11 2009 Page 115 2 Migration to shared raw copper full loop from BSA with PSTN visit and transport to the end customer is not necessary 3 Migration to raw copper sub loop from BSA identical to new installation
132. man resources Information management Legal Procurement Other general administration Interconnection specific costs Subscriber related costs Billing and billing systems 3 2 The treatment of common costs There are various approaches for dealing with common costs Mark ups can be either additive or multiplicative and similarly be either differentiated or uniform A uniform additive mark up implies that common costs are divided by the number of increments and the resultant total is added to each increment thus 1f common costs were 2 000 1 000 would be added to both the core and access increments An additive mark up implies that the allocation of common costs is independent of the total costs of the various increments A multiplicative mark up implies that common costs are split in relation to the relative level of incremental costs of each of the increments For example if the Incremental Cost of Access is 75 of total incremental costs and the Incremental Cost of Core is 25 of these costs then the access increment would be allocated 75 of common Date 30 11 2009 Page 18 costs and the core increment 25 of these common costs Multiplicative mark ups are sometimes referred to as equi proportionate mark ups More sophisticated forms of mark ups are also possible However these methods can be information sensitive and be subject to subjective judgements For that reason mainly equi proportionate
133. mix which would be installed if the network was being built now based on the Modern Equivalent Asset I Cost of Nodes This spreadsheet contains cost information unit costs equipment and installation price trend and asset lives for cabinets including boards within cabinets and NTPs Costs for active street cabinets are shown separately The sheet also contains inputs for economic depreciation parameters I Cost of Links This spreadsheet contains cost information unit costs equipment and installation price trend and asset lives on the following network elements gt Trench and duct gt Miniduct trench and duct gt Copper cable gt Fibre cable Section 3 Cost of copper cable per km allocates the number of copper pair km between SDP NTP ASM PDP and PDP SDP cables The underlying assumption made is that it is the higher cable sizes that are used for ASM PDP cables and the lower sizes for PDP SDP or SDP NTP cables The entries in the cross over points ie the cable size used for both was calculated by use of the goal seeking function in Excel The sheet also contains inputs for economic depreciation parameters I Opex Costs This sheet allows the user to specify a mark up for annual operating cost as a percentage of capex equipment or installation for major cost categories such as trench and duct copper fibre and NTPs Opex costs for some types of assets are calculated in two stages gt Events How often does someth
134. multiplicative mark ups EPMU are used in regulatory cost models In a few cases additive mark ups are used In the current LRAIC Hybrid Model the EPMU method is used for non network related common costs and wholesale costs whereas network related common costs and shared costs are allocated on the basis of cost causality 3 3 Unit costs One of the most challenging tasks when developing a Hybrid Model is to collect robust cost estimates This section reviews the data sources and volumes used and discusses the approach taken to populate the Hybrid Model with cost information 3 3 1 Sources of cost estimates The cost estimates used in the Hybrid Model are based on a number of sources including TDC the LRAIC Working Group various equipment suppliers and benchmarking data from comparable models such as the Swedish LRIC Hybrid Model plus publicly available data sources including Statistics Denmark 3 3 2 Cost estimates in the Hybrid Model According to NITA the cost data used in the Hybrid Model should to the extent possible be recent robust and relevant to a national network in Denmark Therefore NITA has obtained cost estimates from the various sources listed above and developed a consolidated estimate based on all the information available However more emphasis or weight has been placed on estimates accompanied by robust and documented evidence The choice of price data used in the revised version of the Hybrid Model has b
135. n down into different component size categories as far as practical Furthermore costs for special products such as IN services are included though for IN services these are included within worksheet I FA_Costs in the consolidation model Indirect costs for security air conditioning power supply etc are defined as common exchange costs The following allocation keys have been applied in order to attribute common costs gt Power supply unit average power consumption is used as an allocation key because the cost category has a distinct cost driver in power consumption Thus the dimensioning of the power supply unit depends on the total power Date 30 11 2009 Page 98 consumption for the modelled equipment in an exchange average kW per type of exchange gt Air conditioning the dimensioning of the heat development in the exchange This is partially dependent on the size of the room in which the equipment is placed and partially the heat developed by the equipment measured as amount of kWh Therefore power consumption per square meter has been used as allocation key average kW m per type of exchange gt Security security systems site preparation and maintenance with regards to site preparation and maintenance it is fair to assume that these are dependent on the demand for space By contrast it is not fair to assume that there is a distinct cost driver for allocation of cost to security guards and security
136. n model directly from the file directory Enable macros when asked unless the file is from an uncertain untrusted source There may be a message asking whether to update links Such updates are performed automatically on recalculation if other parts of the model are open otherwise links should be updated if other models have been changed and are located in the same directory There is an introductory message that warns the user that older versions of Excel may not provide full functionality 11 4 4 Understanding the model The main inputs in the Consolidation model are not normally altered these are linked to the individual source models However the model contains original inputs relating to functional area costs and staffing options as well as parameters for annualisation and general model setting such as language base year and costing year It also contains the inputs for cost of capital and the NGA selector The recommended approach to gaining an understanding of the model is to work through each sheet in turn starting with inputs moving to calculations and then outputs One way to improve understanding is to edit the input data recalculate and investigate the effect of the changes on the results These can be easily checked in the sensitivity analysis sheet The final costs for all Access Core and Co location products are shown in the output and results sheets of the Consolidation model A sheet by sheet description of the
137. nd moreover these are solely related to the specific co location service The cost categories typically consist of material and or time spent on installation and order processing 10 1 2 2 Demand Demand data has been estimated by NITA on the basis of data provided by TDC The data is updated on a yearly basis The demand volumes for all services are used to calculate the total costs for each service as totals are used as output for the consolidation model Hence the co location model includes specific input cells for demand volumes for each service Date 30 11 2009 Page 106 10 1 3 Cost categories 10 1 3 1 Building and land The costs for indoor and outdoor co location space are calculated in the core model and split into two subcategories internal and external The main reason for this is that land and buildings have different asset lives and price trends Another reason is the need for modelling outdoor co location space Building costs are based on public valuations of TDC s buildings adjusted to reflect current market trends The methodology for evaluating building costs is described in detail in Section 3 5 The estimated costs per m for building and land for indoor and outdoor co location are shown in the table below Table 10 1 Indoor and Outdoor Co location Space excl Mark up Geotype Annual cost per indoor m Annual cost per outdoor m City DKK 765 DKK 210 Urban DKK 463 DKK 33 Rural A DKK 361 DKK
138. nd duct costs used by both networks defined as common costs Costs that are shared by a number of services using a network will comprise a significant proportion of an operator s cost base The difference between shared and common costs is explained below 2 1 1 Shared and common costs Common costs are defined here as the costs of those inputs necessary to produce one or more services in two or more increments where it is not possible to identify the extent to which a specific increment causes the cost This is in contrast to shared costs which are defined here as the costs necessary to produce one or more services within a single increment Box describes the relationship between directly attributable shared and common costs The first definition of the increment as outlined in the previous section would only include some of the directly attributable costs in the core and access networks The second definition would include all directly attributable costs But the variant of the third definition taken to be the increment in LRAIC would include all directly attributable and shared costs in the core and access networks Only common costs would be excluded There are two main types of common costs between access and core that need to be considered First there are the costs associated with trenching and ducting which is shared by the access and core networks By implication any knock on costs associated with this trenching e g maintenance
139. ne the shared costs Allocation is based on the percentage of shared length of trench and a percentage of cost sharing for the shared length of trench which is an input in the I Technical sheet Costs for land based access transmission is calculated directly in the access model As a consequence these costs have been removed in the core model in order to avoid double counting 9 8 3 4 Allocation of shared co location costs There is an element of sharing of site costs with other operators when sites are co located Therefore these shared site costs including sharing of space and air conditioning are allocated to co location 9 8 4 Calculation of service costs Service costs are calculated in the consolidation model The use of network elements has been determined for a given service by multiplying the utilisation of each network element with the traffic volume in busy hour A summation of the utilisation for all services results in the total utilisation of the network element in busy hour This represents the minimum network capacity and is used for network dimensioning as described above Each service s share of minutes in busy hour of minutes for all services have been applied as allocation key for allocating costs to each services Due to the fact that the busy hour conversion is the same for all services this allocation key corresponds to using the service s average usage of the network element divided by the total volume in
140. nect since higher level interconnect is no longer price regulated Additional products have been added to Colocation Costs These encompass additional power variants 6 3 Other LRAIC service costs O_Output Additional products have been added to Full Loops and Sub Loops to cater for a number of additional activities such as unproductive visits and fault handling Date 30 11 2009 Page 50 Additional products have been added to BSA to cater for a number of additional activities such as conversions migrations and changes of speed Additional products have been added to Other to cater for preselection related activities O_Results Additional products have been added to Full Loops and Sub Loops to cater for a number of additional activities such as unproductive visits and fault handling Additional products have been added to BSA to cater for a number of additional activities such as conversions migrations and changes of speed Additional products have been added to Other to cater for preselection related activities 6 4 Co location services O_Output Additional products have been added to to cater for a multiplexing and power products O_Results Additional products have been added to to cater for a multiplexing and power products Date 30 11 2009 Page 51 7 1 Introduction The purpose of this chapter is to provide some guidelines on how to use the LRAIC Hybrid Model Before using the four Excel workbooks that make
141. network was the addition of the data equipment required for BSA products and services up to but not including the Layer 3 and above rings Thus specific equipment was added to the model DSLAMs and Layer 2 Ethernet switches and additional analysis sections added to cater for the transport network i e trenches ducts and cables to interconnect this equipment The following sections provide a more detailed overview of the model modifications relating to the introduction of sub loops and BSA products and services 5 3 Sub loops Sub loop unbundling involves the interconnection of an alternative operator s OAO s network with the local loop infrastructure of TDC at a point somewhere between the ASM Local Remote Concentrator and the end user s NTP Network Termination Point The most likely reason that an OAO would wish to unbundle a sub loop rather than the full loops available at the ASM is to significantly shorten the overall copper loop distance This will then allow the OAO to offer much higher speed xDSL services to their subscribers than might otherwise be feasible over longer full loops Indeed this is also the reason that TDC is currently deploying DSLAMs in various Street Cabinets around the country as they have a declared intention to be able to offer very high speed aDSL services to a large proportion of the country within the next few years Figure 5 1 Illustration of full loops and sub loops SDP ASM j
142. ng capital is calculated differently for Core and Access to show the difference in payment method for the usage of net services This variation is taken into account through different assumptions for the number of debtor days The model assumes that the average number of debtor days is 15 for Access and 105 for Core This difference highlights the difference in payment methods Raw copper is prepaid quarterly whereas interconnection services in the overall net are paid after the end of a quarter billed after being registered Mark up for working capital is added to mark up for other indirect costs to give a total overhead cost mark up The estimates in the model are shown in Table 3 8 Table 3 8 Estimates for Working Capital used in the Hybrid Model Type of cost Estimate Core services 1 17 Access services 0 58 3 5 Building costs for exchanges and technical houses To ensure that the value of the property used for housing equipment such as exchanges and technical houses reflects the real costs NITA has developed a method to assess property values that is more precise than the direct use of public valuations Using this method the public property values as informed by TDC have been adjusted by a factor representing the difference between the market price and the public valuation in each geographical area 3 5 1 Calculation of the real property costs TDC has supplied a set of data with the property records includ
143. ng costs per m and the average land costs per m for each geotype gt The market value of buildings in each geotype is calculated as the average building costs minus the average land costs gt Annual costs per m for land and buildings respectively based on parameters such as asset life salvage value and price trends gt Land costs per m are allocated to outdoor area in m in a 1 1 ratio for each geotype gt Land costs per m are allocated to indoor property costs per m in relation to land area per building area for each geotype Co location occurs at exchanges and not in technical houses making it necessary to calculate a separate price per m for co location which excludes technical houses Technical houses are included in the price per m for the rest of the model In the revised model a number of additional adjustments had to be made to the building values analysis in order to derive updates for 2008 As of 1 January 2007 the county structure in Denmark was reformed whereby the former 13 Amter were reorganised into 5 Regions These in turn are subdivided into 11 Landsdele or subregions In order to derive a continuous data set for market values from Denmark Statistics the old Amter had to be mapped to the new Landsdele This was done by reallocating municipalities from the old Amt areas to the new Landsdel areas based on a transition table provided by Denmark Statistics Adjustments also h
144. node assumption which is explained below Finally it must be decided which specific services should be included in the LRAIC price estimation This is also discussed below 2 1 Defining the increment In principle there are an infinite number of different sized increments that could be measured However these increments can effectively be grouped into three different categories gt A small change in the volume of a particular service gt The addition of an entire service gt The addition of an entire group of services The first definition of the increment refers to the economic concept of marginal costing i e it measures the cost associated with adding a single unit of output In the case of a fixed network a single unit such as a call minute or byte of data traffic is so small that the costs of providing it are effectively zero This makes the approach impractical for modelling purposes and makes it necessary to increase the size of the increment The second definition may apply to services of very different sizes such as interconnection local calls and premium rate calls e g 70 80 and 90 numbers This definition may also be referred to as service based LRIC The Hybrid Model adopts a variant of the third definition The two main increments modelled are the sum of all services in the access network the access increment and all services in the core network the core increment These are described in detail in Ch
145. not including trench sharing does not necessarily lead to a shorter average distance and NITA decided to adopt the second approach because it was more consistent with the constraints likely to be faced by an operator laying down mini ducts in Denmark The average length of mini duct NTP to SDP by geotype derived from this process is shown in the table below gt Raster maps are maps from TDC showing the existing secondary access network It is possible from these maps to measure the distance from the subscriber NTP to the road to the EFSD 4 Tn some cases it was not possible to select 10 houses Date 30 11 2009 Page 64 Table 8 4 Average length of mini duct by Geotype in the hybrid model Geotype Average length per site m City Storby 5 1 Urban By 8 5 Rural A Land A 15 4 Rural B Land B 10 9 Date 30 11 2009 Page 65 8 4 3 Copper and fibre equipment This section describes the methodology used to estimate the amount of copper cables fibre cables street cabinets and network termination points in the hybrid model 8 4 3 1 Overview For the access network the model allows for a three layer tree and branch structure ASM Primary distribution point PDP PDP Secondary distribution point SDP and SDP network termination point NTP The primary input to these calculations is Geographical Information System GIS data for a sample of 20 MDF areas This sample should be constructed to b
146. nput Calculations Output _Costs _Resource O_ Consolidation _ Demand The user can click on any of the buttons on the map to jump straight to the relevant worksheet Each subsidiary worksheet has a Model Overview button which the user can click on to return to the navigation map Most sheets also have a drop down menu below the Model Overview button to take the user to different tables contained in each worksheet Note that it is still possible to navigate the model using the sheet tabs Date 30 11 2009 Page 119 10 3 3 Start up There are two ways to open the Co location model The first method which is recommended is to first open the Consolidation model and then select the Open Linked Sheets option which will open all the individual worksheets of the Hybrid Model This ensures that linked information is updated automatically and links are preserved when the model is saved under a different filename The user should also be aware that the Excel auditing tools will trace precedent and dependent cells only for workbooks that are currently open Alternatively the user may open the Co location model directly from the file directory Enable macros when asked unless the file is from an uncertain untrusted source There may be a message asking whether to update links Such updates are performed automatically on recalculation if other parts of the model are open otherwise links should be updated if other models have be
147. ntinue in the future In general NITA believes that a historical period of no longer than five years is a good starting point for an assessment of the future price trends With regard to estimating the price trends which best reflect the expectations to the future price development a combination of several sources has been used The average price development of the contractors list prices price indices from Statistics Denmark input from TDC and the LRAIC Working Group as well as international benchmarks when relevant In this respect it has been a prerequisite that the development in list prices and documented sales prices all other things being equal is identical whereby drops in prices or increases in prices have the same effect on both sets of prices It has been necessary in a few cases when assessing price trends to apply multiple cost drivers to the relevant price of equipment In these instances the price trend has been based on a weighted assessment of relevant and accessible cost drivers e g hardware software and wages The price trends used in the revised Hybrid Model together with previous values for comparison where relevant are shown in Table 4 3 below Date 30 11 2009 Page 40 Table 4 3 Price trends Cost category Hybrid Hybrid Model Model v2 5 v2 4 Access duct 3 3 Access trench 3 3 Copper cable 6
148. o conducted a simple cross check of its methodology by reviewing maps and considering other relevant data NITA has reviewed the DAV database of roads which showed that there are about 600 000 road segments corresponding to 100 000 roads in Denmark Finally NITA has taken into account that for active street cabinets which are assumed to be sighted alongside the PDPs a compromise needs to be reached between the desire to shorten the length of the copper loop and the need to keep a suitable number of potential customers at each PDP location NITA has therefore chosen to apply the following the number of cabinets in the hybrid mode being sufficiently flexible for an access network provider The numbers shown in the following table are derived from GIS pilot data Table 8 5 Number of SDPs and PDPs in the access network Total number of PDPs in the network 13 800 Total number of SDPs in the network 431 551 Average number of connected lines per PDP 173 Average number of working lines per SDP 5 7 The consequences of this assumption are shown below Table 8 6 Average number of PDPs per MDF area Geotype 1 62 5 Geotype 2 18 4 Geotype 3 9 2 Geotype 4 7 1 Whole network 11 7 Total number of PDPs in the network 13 800 Date 30 11 2009 Page 70 Table 8 7 Average number of SDPs per MDF area Geotype 1 1 731 Geotype 2 873 Geotype 3 255 Geotype 4 126 Whole network 365 Tot
149. od by additional reference to the Excel Comments and by reading comments fields Use of the auditing toolbar functions is highly recommended to understand the model workings Due to the complexity of the Hybrid Model it is not possible to provide details of all functions in the documentation there are far too many cells to consider Each sheet is divided into a series of Tables numbered 1 2 3 etc one below the other Some of these Tables are divided further into sub tables e g 3 1 3 2 etc This means that some worksheets can have many functions and many rows but the functional areas are still clearly delineated Generally data will flow from left to right and top to bottom within a worksheet Drop down menus at the top of most worksheets enable the user to navigate easily through the different tables 7 3 How the models are linked The consolidation model has links from all of the other three models The other three models also have links between them to enable data to flow automatically between the models Each of the three source models Access Core and Co location has an interface sheet Z Interface showing inputs from other models in tan colour coding together with a note indicating the original source of the input This allows users to locate the original yellow coded input cell which can be modified if required Date 30 11 2009 Page 55 Wherever practical data that flows into these three source models from another mod
150. odel is used whereby the costs are first transformed or allocated to network elements and then to services using the traditional routing table method For this the model uses an allocation table and the operational costs that are already allocated to network elements by use of mark ups and in the access model the event driven method The allocation table is comprised of the values of zero do not allocate costs and one allocate costs By using this table and the costs that are already allocated to each network element the model calculates FA costs for each network element The formula used is Date 30 11 2009 Page 25 ij FA x NE xa rA YA where 7 DINE XQ J gt FA FA costs for network element j gt FA FA costs for area i gt NE operational cost allocated to network element j by use of mark ups and in the access model the event driven method gt aj allocation scale for network element j and area i In the model this transformation is developed in the consolidation model in the C FA Costs sheet 3 4 2 2 Allocation of interconnection specific costs Some costs relate specifically to the exchange of traffic or wholesale access services According to Table 3 2 above these costs consist of gt Customer related costs gt Billing and billing systems gt Debtor administration gt Other interconnection specific costs Even though these costs constitute a relatively modest part of the modelled
151. on 1_Line Card 12 Other Subs Layer 2 Aggregation 2 Processor 1_ All Mpps Layer 3 Edge 1 Line Card 1 Layer 2 Agg Gbps Gbps Layer 3 Edge 1 Line Card 2 Layer 2 Agg Mpps Mpps Layer 3 Edge 1 Line Card 14 Layer 2 Bitstream Gbps Gbps Layer 3 Edge 1 Line Card 2 Layer 2 Bitstream Mpps Mpps Layer 3 Edge 1 Line Card 3 Customer Gbps Layer 3 Edge 0 Spare 0 All Gbps Layer 3 Edge 0 Spare 0 All Gbps Layer 3 Edge 1 Line Card 4 L3 Edge Ring Gbps Gbps Layer 3 Edge 1 Line Card 5 L3 Edge Ring Mpps Mpps Layer 3 Edge 1 Line Card 6 L3 Dist Uplink Gbps Gbps Layer 3 Edge 1 Line Card 7 L3 Dist Uplink Mpps Mpps Layer 3 Edge 1 Line Card 8 Other Subs Layer 3 Edge 2 Processor 1_ All Mpps Layer 3 Edge 0 Spare 0 All Mpps Layer 3 Distribution 1 Line Card 1 L3 Edge Gbps Gbps Layer 3 Distribution Line Card 2 L3 Edge Mpps Mpps Layer 3 Distribution 1 Line Card 3 L3 Core Uplink Gbps Gbps Layer 3 Distribution Line Card 4 L3 Core Uplink Mpps Mpps Layer 3 Distribution 1 Line Card 5 Other Subs Layer 3 Distribution 0 Spare 0 All Subs Layer 3 Distribution 0 Spare 0 All Subs Layer 3 Distribution 2 Processor 1 All Mpps Layer 3 Core 1 Line Card 1_ All Gbps Layer 3 Core 2 Processor 1 All Mpps Layer 3 Core 0 Spare 0 All Mpps Peering Router 1__Line Card 1 TDC Side Gbps 1_ Line Card 2 Customer side Gbps Peering Router 2 Processor 1 All Mpps TDM Gateway 14 Media Gateway Controller 1 All Call TDM Gateway 2 Media Gateway 4 Distribution Min TDM Gateway 2 Media Gateway
152. onnect the Active Street Cabinets back to the historical exchange sites This is calculated in worksheet C_Trench by zone using the assumed number of Active Street Cabinets allocated across the 20 sample zones together with a number of other input assumptions that are used to allocate the shared trench and duct systems Under normal running conditions the model does not adopt fibre in the access network to serve existing copper connections thus existing copper supplied subscribers PSTN etc are assumed to be supplied via a copper network The exception to this is where the NGA Selector is switched on In such a case the copper access network is assumed to no longer exist and all customers are fed by a fibre cable In this scenario every PDP site is assumed to contain an Active Street Cabinet and from these locations there is assumed to be a direct point to point fibre feed to the home The model does not therefore currently consider a Passive Optic Network or PON architecture Date 30 11 2009 Page 74 8 5 Treatment of shared costs in the Access Model Most of the cost categories modelled in the access network are shared between different services The costs of all these elements are apportioned to the services that share the infrastructure through the Access Routing table The main inputs to estimate the allocation keys for the different services using the access network are the average number of copper pairs used as means of transp
153. ork used to drive the estimates for the amount of trench and duct required Sections 1 2 contain the size of the road network in km based on inputs from the DAV database and define the amount of trench km per road km by roadtype Then the propensity of different constructions to share routes by different terrain type is defined together with the proportion of the mix of ducts used in different terrain types I GIS Zones This spreadsheet is by far the most complicated of the Input spreadsheets It mainly contains information used to work out distances between the different parts of the access network from RCU to PDP from PDP to SDP from SDP to NTP These are Date 30 11 2009 Page 135 then used to work out copper requirements This spreadsheet is structured according to the 20 MDF areas that have been selected in the bottom up model as sample areas to estimate copper requirements It contains data in Section 3 on the number of NTPs in each sample area and over the whole network as well as number of PDPs and SDPs in Section 6 The section now also allows for an alternative set of PDP quantities under an NGN scenario If inputs in this sheet are changed the macro in C_Cables and Nodes must be re run I Fibre This sheet collects data on the amount of fibre in the network by fibre size and geotype Table 1 contains a calibration scenario for both rings and spurs while Table 4 shows an actual scenario The sheet calculates the fibre
154. ort by the individual services The list of services using the access network therefore includes a category called Others that includes all those services whose costs are not specifically estimated by the model this would require the inclusion of a set of service specific costs on the basis of the total number of copper pairs these use and then assume that this fictional service requires one copper pair Moreover some components of the core network equipment are used by the access related products These costs are calculated in the core model and the costs that are relevant to the access products are transferred across in the Consolidation model These costs include line cards the MDF and some building costs the access part of the core equipment takes some building costs due to the area occupied and also some shared common building costs such as site security and power systems etc Finally some shared common business costs are allocated using an uplift technique equal mark up approach 8 6 Next Generation Access The model has been modified to allow for an exploring of the potential cost impact of moving to a full fibre access network a Next Generation Access or NGA scenario The intention at this stage is that the model will be able to produce indicative results but not at this stage results for use in any regulatory decisions The broad design guidelines currently adopted for the NGA scenario are gt All PDP locat
155. osts are based on the costs calculated in Table 4 5 in the worksheet spread across a number of business areas using inputs contained within this sub section The NP IN mark up is based solely on inputs in this sub section related to the costs of an IN platform It is likely that in future revisions to the model this section will be replaced with specific costs calculated within the core model itself 11 3 Calculation of service costs The final calculation of the costs of the modelled products and services takes place in worksheet C_Services This worksheet comprises three major sections 1 Annualised costs by Network Element 2 Basic product costs Date 30 11 2009 Page 123 3 Final product costs including mark ups 11 3 1 Annualised costs by Network Element This section comprises two parts gt Core and co location network elements gt Access network elements Each part collates the cost data from worksheet C_Costs and presents it in aggregated form by network element The core network element part also handles the allocation of common site costs across the core and co location network elements These allocations utilise input tables in this worksheet to determine which network elements should take a portion of which common site costs These inputs should only be altered if the user fully understands the relevance of these allocations However should additional network elements be added to the model and this part of worksheet
156. other hand a key issue is whether it is appropriate to use the top down assumptions on cable distribution in a world where other elements of the top down approach are violated NITA has identified two important factors gt Different utilisation rates may have an impact on distributional assumptions with higher utilisation rates implying thinner cables In order to correct for this difference between the hybrid and the top down model the cable distribution assumed for the hybrid model should be shifted to the left Total connection length for each part of the distribution network i e SDP to PDP and PDP to RCV is calculated as the product of the average length of each connection i e the distance between SDP and PDP and PDP and RCV and the number of connections this depends on the number of SDPs and PDPs assumed in the area 6 Connection size for each part of the network is calculated as the number of subscriber lines indirectly linked to the point of connection closer to the RCU in the network hierarchy Date 30 11 2009 Page 73 gt Differences in number of PDPs and SDPs may have an impact on distributional assumptions but this also depends on the number of joint boxes that do not enter the model explicitly A smaller number of cabinets would generally imply a higher degree of route sharing and therefore thicker cables While it is not possible to verify and assess the significance of the second factor as sp
157. overed via the recurring fee Thus the costs included in these products are limited to end customer specific administrative tasks and to time charges for an engineer visit where appropriate Date 30 11 2009 Page 116 10 2 5 Pre selection NITA has identified the following relevant tasks connected to establishing pre selection for a new customer gt Processing of order includes all costs attached to pre selection including processing of order confirmation and registration of order It also covers the costs related to the implementation of pre selection The process is assumed to be fully automatic gt Customer service covers the cost arising from customers calling 147 fault complaints A mark up is added to these costs to cover overhead e g IT systems 10 2 6 Implementation of prefix for pre selection When implementing or changing a operator prefix for pre selection a charge per exchange is applied The following tasks is relevant for implementing prefix gt Processing of order includes all costs attached to operator prefix including processing of order confirmation and registration of order gt Technician covers the cost related to the implementation of operator prefix 10 2 7 Interconnection The hybrid model draws on the main activities linked to the different types of cost of points of interconnect Furthermore the hybrid model shows both the costs of installing points of interconnect and their
158. owever for ease of modelling line cards have been included in the core model rather than the access model The costs associated with these items are then allocated to the relevant access products in the consolidation model In this sense it is important to realise that the access network is not the same as the access increment A schematic representation of TDC s access network is provided in Figure 8 1 below Figure 8 1 Schematic outline of Local Access Network Customer Site Ducting Trench Road Assets within the access network include gt the final drop wire to the customer s premises although the cost associated with this drop wire or its activation might be captured through the connection charge gt the trenching in some cases ducted between the final connection point and the remote or host concentrator DSLAM gt radio systems cable and optical fibre in this part of the network Date 30 11 2009 Page 57 gt other assets such as manholes poles and overhead cables if used and gt line cards in the concentrators DSLAMs 8 2 Overview of the Hybrid Access Model 8 2 1 General Approach NITA s preferred approach to modelling access is described in the bottom up MRP and has been used in the hybrid model The preferred approach involves the following stages gt selec
159. p to the Super Core level Each Distribution router site will also contain one or more Edge routers that head logical Edge router rings Every head Edge router will connect to two Distribution routers one at the same physical site and the other at a remote site again for reasons of resilience This is illustrated in Figure 9 3 In version 2 5 of the model the ASM sites have been replaced with DSLAM MSAN sites an MSAN or Multi Service Access Node essentially comprising a DSLAM with POTS and or ISDN line cards Although in theory an operator could combine xDSL customers with POTS ISDN telephony customers in the same DSLAM MSAN the model keeps the two services separate for modelling convenience reasons as much as anything The advantage of this from a modelling point of view is that it improves clarity in the cost allocation process From a technical point of view it might be argued that it also eases segregation of real time voice telephony traffic from non real time xDSL traffic though the existence of a significant number of VoIP subscribers in today s TDC network would tend to reduce the strength of this argument For medium to large sites the adopted approach is not likely to have any real impact on the overall level of investment whereas for the smallest sites it will tend to overstate the number of DSLAM MSAN chassis Overall though the impact will be minor Date 30 11 2009 Page 82 Layer 3 Supe
160. per and in span 48 fibre 10 1 4 2 Overhead Besides the directly modelled costs NITA has applied a mark up to all co location services to take into account the relevant overhead costs not modelled explicitly The mark up is not part of the co location model but is added to the cost of each service in the consolidation model Overhead costs are described in Chapter 3 of this report For co location it is assumed that the net cost of working capital is zero This is due to the fact that the price for co location per m2 is paid in advance whereas consumption is paid in arrears As for the one off costs and costs of establishment caused by co location TDC can ask for payment up front It is assumed that in total these items balance each other out 10 2 Other services in the Hybrid Model As mentioned previously some of the services under LRAIC regulation are not modelled in the hybrid access or core models This is because the services make use of neither the access nor the core network and or they are services related to very specific work processes This section describes how NITA has estimated the costs of these services from a pure bottom up perspective with focus on the specific activities relevant for the different services The services include Raw copper and shared raw copper full loop and sub loop New installation Unassisted New installation Engineer assisted Reactivated installation Unassisted Reactivate
161. pplied to either the calculated numbers or the alternate set of numbers The next two sub sections cover ongoing network costs and one off network costs respectively and allocate the calculated numbers of FTE between Academic Technical and Administrative persons according to the percentages in Table 2 2 of the worksheet The two sub sections also contain inputs for the user to specify the amount of equipment and materials consumed each year by that functional area The final two sub sections cover non network costs and IC specific amp commercial costs respectively and follow the same format as the previous two sub sections although here the data comprises actual inputs rather than calculations 11 2 2 Staff costs The staff costs section simply contains inputs relating to the full annual cost including social costs etc associated with each of the three staff types Academic Technical and Administrative Also contained in this section as mentioned previously are the percentage allocations used to allocate the calculated FTE totals for each functional area to the three staff types 11 2 3 Calculation of pay costs This section comprises of four sub sections covering ongoing network costs one off network costs non network costs and IC specific amp commercial costgs Within each sub section the model calculates for each functional area the pay costs associated with each staff type and then the total for all four staff types Date
162. prices when only the latter have been available gt As an empty entry in the model when only the sum of equipment and installation costs has been available 3 3 3 Data input and confidentiality NITA takes the view that the cost data used in the Hybrid Model should to the extent possible be transparent and accessible for all parties in the process Transparency in the data input makes it possible for all parties to use the Hybrid Model to comment on the data input used and to supply NITA with further estimates and documentation that could enhance the quality of the LRAIC Hybrid Model For these reasons the data input on unit costs e g equipment prices used in the Hybrid Model is whenever possible transparent and accessible to all parties involved in the process Nevertheless in some instances it is necessary to ensure the confidentiality of the data on unit costs This is done using the following procedures gt The method for evaluating the data or their sources is not stated explicitly This means that it is not possible to identify the implicit weight of different inputs used by NITA and thereby to get access to the underlying confidential data gt Moreover when the cost data used are based on confidential data NITA has estimated the cost for a particular type of equipment where cost information is available from several suppliers This is a continuation of the procedure that was used for the first draft v
163. product have been incorporated into the revised Hybrid Model 10 1 4 1 Cables NITA has separately modelled the costs related to materials and installation of the different cabling services 100 and 300 pair copper 8 and 24 pair coax plus in span fibre For each of the cabling services the following parameters are defined gt Cable increment to allow for wastage during installation vV Cable price DKK m gt Digging DKK m if any at all gt Administration order handling etc hours gt Technician physical installation hours gt Other installation costs e g LSA With the assumed cost per hour for administrative and technical staff it is possible to estimate the total installation costs for cables and the cost of the cables It is assumed that the installation cost is independent of the length of the cable whereas the cost of the actual cable is dependent upon the length in 10m increments In addition it is assumed that there are no annual costs Where possible the prices of cables and digging from the access model are used as cost inputs e g this is not possible for coax 8 and 24 pair cables Thus the model contains links to information in the access model via the I_ Interface sheet Date 30 11 2009 Page 109 The co location model computes the following outputs for cable costs the cost per 10m cable and a one off cost for the cable types 8 pair coax 24 pair coax 100 pair copper 300 pair cop
164. r Core Layer 3 Distribution Layer 3 Edge In the Hybrid Model it is assumed that each DSLAM MSAN site will also contain a Layer 2 Aggregation switch This will allow grooming of the traffic from multiple DSLAMSs to a single uplink stream and also facilitate direct connection to customers and or OAOs that require this The Layer 2 Aggregation switches are connected together into logical rings with each ring having two heads Layer 3 Edge router sites This is illustrated in Figure 9 4 Date 30 11 2009 Page 83 rocess DSLAM sa LLU l MSAN S Customer Line Cards Customer Line Cards The model also contains a simple graphical representation of the network structure that was added to aid routing factor analysis This is shown below in Figure 9 5 Date 30 11 2009 Page 84 Figure 9 5 Network Structure Route Chart Layer 3 Core Layer 3 Distribution Layer 2 Aggregation Layer 3 Edge DSLAM MSAN Date 30 11 2009 Page 85 9 3 Network dimensioning 9 3 1 Basic assumptions The main source of information on the current level of demand for the hybrid model is TDC The model includes all the current traffic including gt Voice data and video traffic gt
165. reet cabinet backhaul These are products that could be utilised by OAOs should they decide to unbundle such a street cabinet location For the sake of completeness and modelling integrity a third backhaul product was introduced to cater for TDC s own xDSL and Bitstream customers The full list of street cabinet backhaul products introduced is gt Street Cabinet backhaul duct gt Street Cabinet backhaul dark fibre gt Street Cabinet backhaul xDSL Since the last of the above bullets represents costs more normally associated with the core network the relevant costs are identified within the model and are then allocated back to core based products 8 4 4 3 Cable size The hybrid model assumes a 2 pair copper cable connecting each NTP to the relevant SDP This implies a utilisation rate of 50 per cent with no route sharing for this part of the network On the other hand the hybrid model does allow for route sharing in the distribution part of the access network from the SDP to the ASM This means that the links connecting SDPs with PDPs and PDPs with the ASM share routes and as a consequence they travel on thicker cables and at lower unit costs than they would do otherwise if route sharing was not taken into account Date 30 11 2009 Page 72 The approach used in the hybrid model to dimension distribution cable consists of the following steps gt For each sample area and for each part of the di
166. rket 4 a number of ancillary products have been incorporated into the revised Hybrid Model These are gt Ancillary products relating to installation of LLU unproductive visit by technician cancellation fee calculation of attenuation on the line unproductive fault handling and migration to full loop gt Ancillary products relating to BSA conversion from ATM BSA to eBSA installation of Multichannel change of speed of BSA migration to the different BSA products gt Multiplexing of terminating segments of leased lines gt Installation of street cabinet backhaul access to duct gt Installation of street cabinet backhaul dark fiber gt Installation of bitstream transport Date 30 11 2009 Page 111 gt Installation of Multicast gt Products relating to carrier pre selection In line with NITA s decision on the transit market Market 10 regional POI are no longer price regulated in accordance with LRAIC and have therefore been removed from the co location model Because of the lack of detail and documentation of the available information NITA has to some extent made an assessment of a number of inputs 10 2 1 General assumptions about hourly rate and overhead cost When estimating the costs of individual services NITA has generally differentiated between two groups of activities those related to order processing and those related to carrying out the order Furthermore when modelling the cost of each s
167. road per kilometer of road distance as follows Date 30 11 2009 Page 61 A More than 40 sites on one side of the road per km gt B From 11 to 40 sites on one side of the road per km C From to 10 sites on one side of the road per km gt D No sites on one side of the road In the hybrid model 16 different types of roads are classified as a combination of the density on both sides of the road AA AB etc i e according to the same classes of number of households per kilometer on either side of the road When reclassifying road lengths the number of households on either side of the entire road rather than the number of households on each segment of the road is weighted against the length of the entire road rather than the length of each segment In order to do this NITA has used the databases from DAV and the totals have been updated applying information from Statistics Denmark The length of each road segment was estimated using the MapInfo software in 2002 The following table shows the allocation of streets by the sixteen different categories the classification by geotype is not shown here for the sake of simplicity Table 8 2 Total number of road kilometres Road type Hybrid model AA 3 193 AB BA 5 013 AC CA 468 AD DA 366 BB 17 424 BC CB 7 831 BD DB 903 cc 63 755 CD DC 5 553 DD 8 202 Total 112 708 8 4 1 1 Conversion factors The hybrid model estimates most o
168. roductive such as an unsuccessful visit by a technician or unsuccessful fault handling Date 30 11 2009 Page 113 It should be noticed that costs related to examination and reservation of available trenches are included in the annual price of raw copper Because shared raw copper by definition is always installed on a copper pair which is already in use the estimated cost only allows a limited amount of time in connection with a visit at the end user Raw copper on the other hand is typically installed on unused capacity and therefore transport to the end user is included In addition transport to unmanned exchanges is included 10 2 3 3 Inquiry Inquiries are billed on an hourly basis and the rate is the LRAIC price for technicians Inquiry concerning raw copper and shared raw copper is the hourly rate for a technician plus overhead which is added in the consolidation model 10 2 3 4 New setup New setup of raw copper and shared raw copper is seen as services that are carried out in extension of installation of raw copper and shared raw copper In this respect NITA has included the following cost driving activities gt Processing of order is estimated to be covered by the ordering of the raw copper or shared raw copper and thus it only contains the extra time spent in connection with the new setup gt Physical coupling in exchange includes coupling in the access network in connection with the installation of raw copp
169. rol sheet is reproduced in Figure 11 2 This user interface enables the user to navigate the Consolidation model The user can click on any of the shortcut buttons on the map to jump straight to the relevant worksheet or table Each subsidiary worksheet has a Model Overview button which the user can click on to return to the navigation map Most sheets also have a drop Date 30 11 2009 Page 125 down menu below the Model Overview button to take the user to different tables contained in each worksheet Note that it is still possible to navigate the model using the sheet tabs Figure 11 2 Consolidation model Control Sheet g T Change la J IT og Telestyrelsen a Mirestenet tor Videnskab Tekasiog og Udvibling Open linked models Model results Navigation sheets Can only be used when other models are open Open linked sheets Show results and have the same version number Input to consolidation Print results Hybrid Core Model From other models Sensitivity Hybrid Access Model Parameters User information Functional areas x Hybrid Co location model Overview Calculations in consolidation Sensitivity analysis Colour codes Cost categories v2 4 2009 resus _ Compare Services Stored results Compare Current results j Compare Functional Areas Store current results Other input Cost of Capital Overheads Date 30 11 2009 Page 126 11 4 3 Start up The user should open the Consolidatio
170. ry The adjustments concern gt the application of routing factors gt the adjustments for grade of service gt allowance for resilience gt consideration of the burstiness of the service and gt application of the busy hour estimate 9 3 1 1 Routing factors The routing factors show the way in which the different network elements are used in the process of establishing a call In an IP network this potentially involves many more network elements than in a traditional PSTN The following example may help to illustrate this A legacy local call can be routed through a remote subscriber stage and then on to a local exchange from which it is directed to another remote subscriber stage In this case the call setup uses Date 30 11 2009 Page 86 gt 2 ASM or remote subscriber stages gt 1 local MSG exchange gt 2 transmission connections between the remote subscriber stage and the local MSG exchange In an IP network by contrast packets may travel through many more switching points Layer 2 switches and Layer 3 routers in the case of an IP network than would typically be the case for a legacy PSTN network As shown in Figure 9 6 a local call made over an IP network might take a complex route involving for example the following Network Elements gt Multi Service Access Node processor at the originating end of the call gt MSAN facing line card in a Layer 2 Ethernet Switch gt Processor
171. s been manually created The intention is to show a potential DWDM network All of the sites links and distances have been manually entered in the graphic based on the information in _Trenching_Transit and also the Autoroute mapping Ringk bing Skjern 24 5 Brondersiev I Hjorring Viborg Herning Skjern Esbjerg Kolding Centrum Stovring 20 1 Byen Hobro 29 1 R nne 170 Randers M Randers 26 1 30 6 Viborg 40 4 Arhus Contrum M Arts conmo a Br nderslev 29 2 23 1 Thisted 94 5 Skive 59 2 Arhus Viborg Centrum Silkeborg ial Silkeborg 36 117 Borups Alle 269 49 47 7 Herning 37 Hasselager 38 9 M Hasselager 16 K ge Albertslund 26 9 Albertslund 46 Skanderborg 18 1 Borup 14 80 41 2 Neks 224 Odder Borup Horsens 26 7 Kors r Kors r 65 9 Nastved 52 8 56 7 Vejle 26 6 Odense 56 2 Odense Vordingborg Nyk bing Roseng rden 4 Svendborg 40 9 Svendborg 100 S nderborg 68 5 Abenra 32 7 The individual routes were then assessed both in isolation and in aggregate to arrive at a compromise network with DWDM links that addressed the 80km limit whilst seeking to minimise the overall level of investment in DWDM equipment In this respect not all of the complete distance for a link greater than 80km was deemed to run over the DWDM network Instead the tails of the link were allowed to remain non DWDM as long as the length of the tails remained within limits 9 6 3 Equipment configuration For pra
172. s that support the model will be disabled and the model will not work properly Running the consolidation model opens a Control Sheet This allows the user to select the model language English or Danish by clicking on the country flag The settings will automatically feed through to all other models It should be noted however that some recent additions to the Hybrid Model are available in English only In order to find service costs the user should click on the Show Results box The O_Results sheet provides a quick overview of final unit costs for each regulated service reproduced from the O_ Output worksheet 7 2 Overview of common model features The models are Excel workbooks Each Excel file has the following logical structure gt Descriptive navigation sheets gt Input worksheets Date 30 11 2009 Page 52 gt Calculation sheets gt Output sheets The name of each input calculation or output sheet is prefixed I_ C_ or O_ respectively to indicate its primary status They also use different coloured tabs yellow for inputs dark grey for calculations and blue for output pink for output to consolidation sheets Light grey tabs indicate descriptive or navigation sheets There are some overlaps between sheet functionality These have been introduced to ease both use and development since they help to avoid a proliferation of worksheets Thus some calculations may be done in input sheets and there may b
173. s to place on another provider s premises one s own exchange and other equipment intended to transmit and control signals between specific termination points in connection with the exchange of traffic lease of infrastructure capacity or service provider access gt Access to carry out one s own operation and maintenance of exchanges and equipment The hybrid model does not estimate the costs of all of the different types of co location LRAIC based costs for co location are limited to costs associated with sharing facilities such as buildings and exchange equipment in accordance with section 6 1 and 7 2 regarding gt agreements on exchange of traffic and gt the lease of non equipped infrastructure sections in the subscriber network The responsibilities and obligations with regard to co location are described in detail in Agreement on Co location between TDC Totallosninger A S and Company X 10 1 1 1 Co location for the exchange of traffic For co location related to exchange of traffic the access seeker rents space on the access provider s premises to interconnect with that operator In the process the access seeker installs its own exchange if needed and its own interconnection equipment The access to a core network for the purpose of exchanging traffic is usually considered under three headings gt Physical co location where the access seeker chooses supplies installs and operates the equipmen
174. se that installation of NTPs does not require installation of raw copper shared raw copper Thus installation of NTPs is billed separately at the price determined by LRAIC and not at the incremental cost of simultaneous installation 10 2 3 2 New Installation Unassisted Engineer assisted Related to the installation of raw copper and shared raw copper for both full loops and sub loops the following cost driving activities should be considered gt Processing of order consists of cost of ordering before and after the physical coupling in the exchange e g reception confirmation and putting the order into the system and is considered an automatic process gt Physical coupling in exchange includes re arrangement of cross thread plus establishment of connection from the wiring board via the MDF gt Transport time to exchange covers the time spent on transportation to and from the exchange when the exchanges are un manned gt Visit at end user contains the examinations of the connection at the end user and potential changes e g internal cross connections in the MDF that are undertaken from the exchange to the end user This activity is only undertaken when installation is assisted by engineer gt Transport time to end user covers the time spent on transport between the exchange and the end user As mentioned earlier the revised Hybrid Model now also includes costs for activities that are ultimately unp
175. send correct signals about the real economic value of the network ensuring stable pricing which is not dependent on the accounting principles chosen by the network owner Correctly applied forward looking replacement costs in the same way as historic costs will also ensure full recovery of the original installation costs of the network neither more nor less Date 30 11 2009 Page 42 4 6 3 Storage and factory buildings and land The price development for storage and factory buildings and land can be found in Table 4 5 below Table 4 5 Changes in the Index for Storage and Factory Buildings and Land Per Annum Industrial buildings Land 2001 3 6 2 2 2002 4 4 2 4 2003 3 0 1 4 2004 4 2 6 6 2005 7 4 4 0 2006 11 6 9 7 2007 15 1 4 9 2008 Q1 Q3 2 4 6 4 Average 2001 2008 4 3 4 7 Source Statistics Denmark The prices of storage and factory buildings have fluctuated considerably in the period of 2001 to 2008 The same development can be identified regarding the prices of land albeit to a lesser extent Given the recent fall in property prices which looks likely to continue for a number of years NITA has decided to use a price trend of 0 for buildings Land prices have been more stable and therefore NITA has based its price trend on the average over the last 8 years rounding down to 4 Date 30 11 2009 Page 43 5 1 Introduction In 2006 NITA extended the scope of the ex
176. sheet it contains some calculations as this increases transparency and eases audit of the calculations Section 1 contains the functional area staffing for network costs non network costs and interconnection specific costs Network costs are subdivided further into ongoing costs and one off costs Within Network Costs inputs are provided for minimum staffing requirements for each functional area plus an additional number of FTE staff that is linked to an opex cost driver as specified in Table 1 1 Users may change the yellow coded staffing inputs but should only modify the relevant opex driver after careful consideration It is also possible to override the default inputs by adding values to the Options column which may be selected in the FA cost option field at the top of the worksheet Note that input data must reflect wholesale costs relevant to modelled services only No retail costs or other non relevant costs related to the core or access network should be included In Sections 2 and 3 the staffing requirements in full time equivalents FTE are then multiplied by the staffing costs for each type of personnel administrative technician and academic professional To derive total functional costs non pay costs for equipment and material are added to pay costs in Section 4 Finally Section 5 provides inputs for overheads and working capital which are used to calculate common cost mark ups I Costs This worksheet contains linked
177. sible to use a universal override if the user wishes to apply a method other than the model default across all cost categories This can be specified in Z Parameters 14 4 3 Calculation sheets C_FA_Costs The sheet is used to allocate the various Functional Area costs across the various network elements Vertically the sheet contains the network elements Horizontally are the FA categories and their costs as calculated in the Z F A_ Costs sheet Column H shows the initial estimate of the operating costs by network elements that is an input from the core and access models Column I sums the equivalent FA costs allocated to network elements In columns K to AK the user may change the allocation keys used The keys can only be 1 or 0 where 1 allocate and 0 don t allocate The cost calculations are performed in columns AM to BM C_ Costs This worksheet performs the final cost calculations on the various cost components and then allocates them to the available network elements C_Services This sheet allocates the network element costs to services and then adds on the relevant mark ups Section 1 is a matrix of the costs allocated to each model by network element and the horizontal axis and cost type on the vertical axis Cost types are broken down into broad categories such as expensed equipment costs expensed installation costs annualised equipment costs annualised installation costs and annualised operating costs For access network elemen
178. solidation model therefore contains information from the other three models which is retrieved through links Date 30 11 2009 Page 124 The overall model architecture is illustrated below Figure 11 1 Model Architecture ot mM mM The Consolidation model calculates the annualised costs of each cost category identified in the three main models The main components of this are gt Calculation of operational and indirect costs gt Calculation of the annualised cost of the capital equipment expense and gt Calculation of the annualised cost of installation costs After calculating the annualised cost of each cost category the model then attributes each of the cost categories to a network element Next by using a routing allocation table for access and core services the model calculates the individual service costs Some costs may be expensed not annualised but are assumed to be paid for with a one off payment These expensed items are also processed as the access and co location service costs may be a mixture of annualised and one off costs Further the consolidation model carries out gt Calculation of working capital gt Allocation of overhead costs 11 4 2 Model structure The model includes a control sheet which provides shortcuts to important parts of the model such as Results It also allows the user to change the model language print results and open other models The UK version of the cont
179. some would result in an under dimensioned network and increased costs for the remaining services as shared costs such as duct would be allocated to fewer services The core and access models categorise services under the three broad headings PSTN including ISDN and broadband including bitstream leased lines and other services described below 2 3 1 Voice and broadband data services Voice services include standard call services that originate and terminate on exchange lines Broadband services include both broadband to own customers and wholesale bitstream services Table 2 1 lists the most important services Date 30 11 2009 Page 15 Table 2 1 Voice and broadband services Core Access Local calls PSTN Line Rental National calls ISDN 2 Line Rental International calls Inbound ISDN 30 Line Rental International calls Outbound Wholesale Line Rental International calls Transit Shared Access Fixed to mobile calls Full Access Mobile to fixed calls Other Access including fibre and wireless technologies IN Basic IN Advanced Broadband Interconnection Transit Within Area Interconnection Transit Between Areas Broadband to own customers Interconnection Local Area Bitstream access Interconnection Within Area Other broadband based services such as VoIP and IPTV Interconnection Between Areas Operator Services Other calls Where the demand for these services has not been included
180. ss network 13 2 Services in the core network 15 0 Co location services 13 8 Other services 15 0 These mark ups are based on calculation of a common mark up for all the services The common mark up was found to be 13 80 percent It has subsequently been adjusted for working capital see the paragraph below for the individual categories This results in four individual different mark ups 3 4 2 4 Estimation of FA costs FA costs in the model are based on an offline analysis of the available data from the existing LRAIC model TDC s latest cost statement TDC s previous top down model and other benchmarking data such as the Swedish LRIC model However these data may not be used directly To consider the actual modelled network and the FA method s cost categories it was necessary to make a series of corrections of the data material available The method is therefore an iterative process consequently both a bottom up and a top down approach have been used with a subsequent calibration The bottom up approach is based on dimensioning factors defined by experience such as number of employees per central or installation time for a certain type of equipment converted into personnel hours This approach is typically used for development of business plans and investment analysis The top down approach is however based on general key numbers such as maintenance s share of GRC After convergence of the results of the two m
181. stribution network cable requirements in terms of length by size is obtained by multiplying total connection length by connection size connection size excludes the adjustment for growth at this stage gt Length by size cable requirements by geotypes are worked out by multiplying length by size requirements for each sample area by the weight given to the area in question gt Allowances for spares are taken into account through inputs that are specific to each part of the network and each geotype This would increase length by size cable requirements gt Length by size cable requirements by geotypes are then allocated to different cable size categories through specific distribution assumptions The distribution assumptions are derived from the top down model It is NITA s view that this method offers the following advantages gt It makes use of intermediate outputs produced by the bottom up model and in particular of the distances between cabinets worked out in the original model in a very detailed way gt As the original reconciliation exercise has shown working on the variable length by size allows separation of the issues of connection lengths connection number and their sizes It therefore enables the problem of cable distribution by size to be dealt with as the last step of the dimensioning process gt All the necessary information is available On the
182. t and any mark ups applied Access costs can be analysed separately by geotype if required Date 30 11 2009 Page 145 15 Appendix D Bilag vedr kortlagte andre tjenester for LRAIC 2010 Date 30 11 2009 Page 146 Date 30 11 2009 Page 147 Date 30 11 2009 Page 148
183. t and cable copper and fibre from the active street cabinet sites to the exchange building Date 30 11 2009 Page 59 After calculating the resources needed to build this part of the network the model uses the results to calculate the capital cost of those resources and the event driven annual operating cost of such a network The methodology used to work out the equipment requirements is summarised in the table below and explained in further detail in the following sections Table 8 1 Summary of methodology used to estimate major network elements in access Network element Summary of methodology Trench and duct The main methodological assumption that has been adopted in the bottom up model in order to model trench requirements in the access network is that trenches in Denmark have a direct relationship with road paths The DAV database has been used to classify all roads in Denmark in the sixteen road categories created and in the four given geotypes For each of the sixteen types of road a factor ranging from 0 4 to 1 8 has been assumed to convert road length into trench length These are referred to as conversion factors Copper Cabinets The methodology adopted to work out copper requirements is quite complex and relies on the results of a dimensioning exercise performed on a sample of twenty MDF areas hereafter called sample selected by the LRAIC Forum among TDC s 1 183 exchange areas in 2
184. t is recommended that this should be done by experienced Excel users only Date 30 11 2009 Page 142 14 4 2 Input sheets I Interface This sheet pulls in inputs from other parts of the model that are used by the Core model These are shown in tan colour coding together with a note indicating the original source of the input Access Core Co location This allows users to locate the original yellow or blue input cell which can be modified if required I Parameters This worksheet contains various parameter information carried across from the three feeder models such as demand volumes and cost drivers In Section 1 all the Network Elements from the different models are listed and summarised A check is performed to ensure that all cost categories have been allocated to a network element Section 4 contains parameters for a range of annualisation methods with an option to enforce use of a single method across all cost categories This will override any annualisation options applied to individual assets in the J_ Costs sheet The model default is economic depreciation which is simulated by a mix of tilted annuities and output adjusted tilted annuities conditional on parameters related to cost structure output trend and price constraints The annualisation method is explained in detail in Chapter 4 of this report I FA_Costs This sheet contains input information used to calculate the Functional Area costs Although this is an input
185. t needed on the premises of the access provider and therefore access has to be provided for the staff of the access seeker Date 30 11 2009 Page 104 gt Managed co location sometimes referred to as virtual co location where the access seeker chooses and supplies the equipment but installation operation and maintenance is carried out by the access provider gt Direct connection where no additional equipment is needed 10 1 1 2 Co location for raw copper With regard to co location in connection with raw copper the access seeker will rent space on the access provider s premises in order to install and operate its own equipment usually xDSL to be used in connection with the local loop This is illustrated in Figure 10 1 Figure 10 1 Co location Network Termination Point NTP Customer Operator Raw copper circuit gt Typically there are two different options gt Physical co location The operator can physically locate its equipment at a site belonging to the SMP operator containing an MDF An internal tie cable supplied by the SMP operator connects the chosen subscriber lines in the MDF with the operator s equipment typically a Digital Subscriber Line Access Multiplexer DSLAM For security reasons it is necessary to keep the location of the two Distribution Frames separated gt Virtual co location In cases where it is not possible for the SMP operator to physically locate equipm
186. the access network was essentially modelled in two ways the trench and ducts were modelled on a national basis whereas the cable was modelled on the basis of 20 sample areas In both cases the emphasis was on calculating the total cost of connections between the ASM local remote concentrator and the NTP subscriber network termination point For version 2 4 the major change implemented with respect to the modelled access network was to analyse the required trench and duct also on the basis of the same 20 sample areas as used for the cable in order to split the corresponding equipment requirements into two parts the primary network from the ASM TH to the PDP and the secondary network from the PDP to the NTP Care was taken to ensure that the resultant national totals were consistent with those previously calculated In the Core part of version 2 3 of the model the network was analysed separately for nodes switch locations switching and transmission covering both the active transmission equipment and the core trench and duct network The emphasis was very much placed on the costs of standard voice products and services Data services leased lines and other services were essentially only considered in sufficient depth to allow for a suitable allocation of common and shared costs such as core trench and duct for example Date 30 11 2009 Page 44 For version 2 4 the major change implemented with respect to the modelled core
187. the most appropriate method are Date 30 11 2009 Page 34 gt Whether output is constant rising or falling gt Whether asset costs are fixed or variable gt Whether there is a price constraint due to market conditions or regulation Different combinations of these conditions can warrant different methods of annualisation as shown in the matrix below Even when Economic Depreciation is selected as the default in practice the default method used in the Hybrid Model is gt Black boxes Tilted annuity gt Red boxes Output adjusted tilted annuity Economic depreciation Date 30 11 2009 Page 35 Change in Output Asset purchase price Fixed costs No advantage gained by later entrant as Variable costs Change in output is not relevant as the to keep a static unit price over the years which would equate to higher depreciation in the early years ane they still have to buy the same quantity extra assets purchased each year are of the asset over time since all costs are fully utilised over their lifetime a 3 fixed i e do not depend on output amp trends 2o If there are market and or regulatory Change in output is not relevant as the constraints on the product market price extra assets purchased each year are D and output is increasing then the fully utilised over their lifetime D depreciation charge should be be 5 increased over time to reflect the higher 5 value of
188. three parts Date 30 11 2009 Page 8 gt Common Guidelines for the Top down and Bottom up Cost Analyses gt Guidelines for the Top down Cost Analysis gt Guidelines for the Bottom up Cost Analysis An updated version of the Model Reference Paper has been published on NITA s website and forms the basis of the current revised version of the Hybrid Model 1 4 The purpose of this report This report sets out to describe the overall methodology underpinning the Hybrid Model It also provides detailed documentation on the different components of the Hybrid Model access core co location and other LRAIC services as well as information on how costs were determined In particular it documents revisions that have been introduced up to version 3 1 of the Hybrid Model In addition the report incorporates a user guide This offers practical guidance to interested parties about how the models work and on the nature and role of each of the spreadsheets A general introduction to the model is provided in Chapter 7 and more specific guidance for each individual model Core Access Colocation and Consolidation is included in the relevant chapters in Sections 8 to 11 Furthermore in relation to the revision of the hybrid model NITA has on 14 November 2008 updated a draft note on the identification of Other services that are included in the co location model In this consultation NITA received comments from TDC the Danish Energy an
189. ting a sample of exchange areas from each geotype gt setting the boundary for each of these exchange areas on the basis of the boundary in the SMP operator s own network gt on the basis of detailed maps determining the optimal layout of a network given the known number of subscribers for that exchange area the dwelling and street pattern gt aggregating up to estimate costs for the geotype as a whole and in turn for Denmark e g if the proportion of subscribers examined in a particular geotype is 10 costs are multiplied by 10 The hybrid model relies on geographical data provided by GIS databases The model calculates the distance of the primary access network i e the part of the subscriber line that runs from the exchange to the primary distribution point generally a street cabinet and distance of the secondary access network i e the part of the subscriber line that runs from the primary distribution point to the final drop to the customer s premises Information is needed for each of these distances separately a to facilitate the costing of sub loops and b because the dimensioning rules for these different parts of the access network will differ For example the slow and erratic growth on the secondary access network makes it much more difficult to dimension than at the aggregated primary access network For this reason provision of secondary access network capacity will usually be more generous than
190. tion equipment The methodology thus seeks to minimise if not remove the effects of outdated technological solutions and inefficient operating practices One important conclusion that NITA arrived at was that there now needs to be a complete updating of the modelled network to one based on packet switching technology NITA believes that prices based on such a network should be incorporated into the price regulation starting from 1 January 2010 accepting that consideration Date 30 11 2009 Page 7 might still need to be taken of the fact that TDC could not be expected to have completed an NGN migration by that time 1 2 The use of LRAIC price estimation methods By using the LRAIC price estimation method the total price for an interconnection product cannot exceed the sum of long run average incremental costs of producing the relevant interconnection product as stated in 14 section 2 of Executive Order no 930 The LRAIC price estimation method consists of a spreadsheet model developed in Microsoft Excel which is published on NITA s website The model named the LRAIC Hybrid Model is the basis for the determination of LRAIC based prices for various interconnection products All material that has previously been part of the development of the LRAIC price estimation method is with due respect to the rules concerning exceptions from the right of access to documents in the Act on Right of Access to Documents Offentlighe
191. transmission link gt A second instance of the Ring facing line card in the Layer 2 Ethernet Switch gt A second instance of the Processor in the Layer 2 Ethernet Switch gt A second instance of the MSAN facing line card in a Layer 2 Ethernet Switch gt Multi Service Access Node processor at the terminating end of the call In the hybrid model the routing factors together with the traffic volumes are used for dimensioning the number of exchanges and transmission capacity it should be noted that demand for transmission capacity is also driven by leased lines and other services Date 30 11 2009 Page 88 Figure 9 6 Routing factors an example J i ili Layer 3 Core Layer 3 Distribution ne Layer 2 Aggregation DSLAM MSAN Sils si e is aa 9 3 2 Method for network dimensioning The dimensioning of equipment is based on the routing factor method whereby traffic data for the individual routing factors are applied to the dimensioning rules of the relevant equipment The model converts call minutes into Busy Hour Erlangs BHE in the sheet I Product Demand and data traffic into Busy Hour Gigabits per Second BH Gbps In the next step a routing factor is applied to define the extent to which the different network elements are used by the different services Date 30 11 2009 Page 89 Each product modelled within the core network contains entries in the routing factor
192. ts costs are further broken down by geotype Date 30 11 2009 Page 144 The remaining sections allocate the network element costs to individual products and services and apply mark ups for common costs to derive a final unit cost per product The results are fed through to the O_Output sheet 14 4 4 Output sheets O_Output This sheet provides a complete set of unit costs for the products and services modelled in Core Access and Co location One off costs are shown separately from annual costs as appropriate For access related costs outputs are shown by geotype as well as for the country as a whole O_Results This is the final results sheet of the Consolidation model It shows a more limited set of outputs than the O_Output sheet focusing on results of particular interest or those which will have a significant impact on pricing decisions O_Sensitivity The sensitivity sheet allows the user to check the sensitivity of any changes made to model inputs It compares current results with previous results and shows percentage differences calculated as current previous The default comparison in the revised hybrid model is with the previous model version Version 2 4 which was first released in 2006 and updated on an annual basis for changes in demand Section 4 of the worksheet provides a disaggregated cost analysis to allow the user to explore the relative importance of the various network elements that make up the unit cost of a produc
193. ts per second as percentage of downstream 80 IPTV parameters Average size of IPTV packets Bytes 1 350 Million pps per Gbps conversion 0 09 Header total 82 Bit Rate uplift due to headers 6 1 Upstream bits per second as percentage of downstream 0 03 Upstream packets per second as percentage of downstream 0 20 non PSTN parameters Average size of non PSTN packets Bytes 750 Million pps per Gbps conversion 0 17 Header total 62 Bit Rate uplift due to headers 8 3 Upstream bits per second as percentage of downstream 100 Upstream packets per second as percentage of downstream 100 Flexibility with regard to the technical parameters of new services is important to enable the model to calculate the network demand resulting from the predicted levels of usage of the various products and services A good example of this would be for VoIP where not only can different codecs be used G 711 G 729 etc but also IP header packets can be compressed and periods of silence removed from the packet flows The selected combination of these parameters can have a significant effect on the resultant demand placed on the network A case in point is the effect of the relative average size of IP packets on cost allocations based on packets per second rather than bits per second This can result in differences in the amount of cost allocated to voice telephony services by a factor of between four and five This highlights the importance of having clear param
194. two or more services within the same increment where it is not possible to identify the extent to which a specific service causes the cost Examples of shared costs in the core network include optical fibre transmission equipment and related overheads all used by voice data and video services Common costs are the costs of those inputs necessary to produce one or more services in two or more increments where it is not possible to identify the extent to which a specific increment causes the cost Trenching costs provide a good example of the difference between shared and common costs The costs of trenching specific to the access network or the core network will generally be a shared costs since the trenching is likely to be used by two or more services However some trenching will be used by both the access and the core network In these instances the costs will be common costs 2 1 2 Defining the core and access increments The advent of new technology is increasingly blurring the boundary between the core and access networks Nevertheless the traditional definitions of core and access described below provide a useful basis for measuring the costs of the two main increments Date 30 11 2009 Page 12 2 1 2 1 Core Costs in the core network are primarily driven by the volume of traffic and by the number of call attempts whereas costs in the access network primarily are driven by the number of subscribers In practice the number o
195. ue background These styles can be summarised as follows gt Input data Yellow boxes indicate an input The user may enter a value or occasionally enter a hard coded formula to create the value required this is not recommended but it can be useful gt User notes Lilac shading indicates important notes or the source of the data Changing these cells will have no effect on results These cells generally describe neighbouring cells gt Input parameter Pale blue boxes indicate that the values could be changed but require thought before doing so This is because changes will alter the meaning of other data or because the values are key parameters that affect a considerable part of the model calculations The inexperienced user should generally leave these unchanged or at least check the consequences e g using the Excel Audit tools before making a change gt Header at start of table Each worksheet has areas tables that relate to a type of calculation or processing of data These areas are separated from each other by a header row blue text on a pale grey background Date 30 11 2009 Page 54 gt Calculations Developer notes Warnings Error messages Unfilled white cells are used for working or for text comments that describe a neighbouring cell These should not be altered unless you are developing the model gt Bold red text or cells shaded pale red indicates warning messages some data has not passed
196. ugh a detailed 30 year forecast may look impressive it relies heavily on assumptions relating to annual output volumes price trends asset replacement cycles and WACC rates which become progressively more spurious the further into the future the model attempts to forecast A simpler approach which has been adopted in the Hybrid Model is to adjust the tilted annuity formula so as to account for changes in output as follows Date 30 11 2009 Page 36 Annuity Purchase Price WACC Adjusted_Price_Trend 1 1 Adj usted _Price_Trend 1 WACC where Adjusted Price_Trend 1 Output_Trend 1 Price_Trend 1 Assuming constant price and output trends this formula gives the same results as a detailed economic depreciation calculation While the method is not able to model annual variations in output volumes or asset prices that do not follow a single trend such detailed and varied forecasts are unlikely to be accurate many years into the future and tend to add little value to the final result particularly as applied to a one year model NITA therefore believes that the adjusted tilted annuity method can be used as a proxy for economic depreciation Operating Expenditure The same output adjusted tilted annuity formula can be applied to operating expenditure As assets get older they generally require more maintenance As a result the operating expenditure associated with network assets tends to increase over time
197. ular care should be taken when changing array formulae including single cell array formulae Adding new features or processing can be simple or complicated depending on the scope The developer must understand the model and the Excel formulae Changing the names of named arrays named cells or the name of a worksheet is highly likely to give errors especially to data links between the models Since the models include a number of macros written in Visual Basic for Applications users should not change any sheet names as this will cause the macros to malfunction Many calculations use lookup formulae based on data labels and headings in another data sheet These should only be amended therefore after careful investigation of the consequences Please note that if a file name is changed which is allowed and indeed this is recommended if there are changes as part of version control policy then Excel links in another model must be updated since the other model will still link to the previous version If two models are open simultaneously then saving one to a new file name will result in the links automatically moving to the saved file It is recommended therefore that all models should be open while files are saved under a new name In addition all files should be individually saved prior to closing down Excel This will ensure that any links between files refer the most recent model version Date 30 11 2009 Page 56 This section sets out
198. unassisted from raw copper full loop identical to new installation unassisted from shared raw copper full loop identical to new installation unassisted from shared raw copper sub loop visit and transport to the end customer is not necessary 4 Migration to shared raw copper sub loop from BSA same as new installation unassisted from raw copper section same as new installation unassisted 10 2 3 11 Unproductive fault handling In case of fault handling and when the technician can not get access to the installation for instance when the end customer are not at home a charge is applied 10 2 4 BSA installation The costs of BSA installation are broken down into the following individual products or services gt New Installation Unassisted gt New Installation Engineer assisted gt Reactivated installation Unassisted gt Reactivated installation Engineer assisted gt Additional work for a new installation of BSA without co production gt Conversion from BSA on line to BSA without co production gt Conversion from ATM BSA to eBSA gt Installation of Multichannel gt Change of speed BSA gt Migration to the different BSA products from BSA without telephony from raw copper full loop from shared raw copper full loop For BSA new installation and re activation NITA decided that the costs related to the DSLAM including the line card should not be allocated to installation but rec
199. us added into the model to allow the user to define among others gt The proportion of xDSL lines fed by Street Cabinet based DSLAMs gt The number of Street Cabinets having DSLAMs gt The number of ASM Technical House locations on Layer 2 rings gt The number of Street Cabinets also populated with smaller Layer 2 Ethernet switches Date 30 11 2009 Page 48 5 5 Ancillary products In addition to the principal products exclusive and shared sub loops and BSA a number of ancillary products were also included in the 2006 model gt Ancillary products relating to sub loop installations installation request new set up physical rearrangement change of trunk number technical assistance gt Ancillary products relating to BSA new installation unassisted and engineer assisted reactivation unassisted and engineer assisted additional fee for installation without co production conversion to BSA without co production new NTP for BSA gt Products relating to co location within a technical house one off and recurring fees for 1 5 rack space installation and recurring fees for 48V power for same For BSA new installation and re activation NITA decided that the costs related to the DSLAM including the line card should not be allocated to installation but recovered via the recurring fee Thus the costs included in these products are limited to end user specific administrative tasks and to time charges for an engineer visit
200. usage is defined in worksheet I Network_Elements and is illustrated in Figure 9 7 Figure 9 7 Network Elements in the Core Model Main Sub Item Units of Usage Spare 1 Spare1 1 All Subs Spare 2 Spare2 1_ All Subs Spare 3 Spare3 1 All Subs FTTH 4 ODF 1_ All Subs DSLAM MSAN 1 MDF 1_ All Subs DSLAM MSAN 0 Spare 0 All Subs DSLAM MSAN 2 Line Card 1 POTS Subs DSLAM MSAN 2 Line Card 2 Adsl Subs DSLAM MSAN 0 Spare 0 All Subs DSLAM MSAN 0 Spare 0 All Subs DSLAMMSAN 2 LineCard 3 Sdsl Subs DSLAMMSAN 2 LineCard 4 Vdsi Subs DSLAMMSAN 3 Processor 1_ Card related POTS Subs DSLAM MSAN 3 Processor 2 Card related xDSL Subs DSLAMMSAN 3 Processor 3 Traffic related POTS Mpps DSLAMMSAN 3 Processor 4 Traffic related xDSL Mpps Layer 2 Aggregation 1 LineCard 1 DSLAM POTS Gbps Gbps Layer 2 Aggregation 1 LineCard 2 DSLAM POTS Mpps Mpps Layer 2 Aggregation 1 LineCard 3 DSLAM xDSL Gbps Gbps Layer 2 Aggregation 1 LineCard 4 DSLAM xDSL Mpps Mpps Layer 2 Aggregation 14 Line Card 5 L3 Uplink Bitstream Gbps Gbps Layer 2 Aggregation 1 Line Card 6 L3 Uplink Bitstream Mpps Mpps Layer 2 Aggregation 1__Line Card 7 Customer Gbps Layer 2 Aggregation 0 Spare 0 All Gbps Layer 2 Aggregation 0 Spare 0 All Gbps Layer 2 Aggregation 1 Line Card 8 L2Ring Gbps Gbps Layer 2 Aggregation 1__Line Card 9 L2Ring Mpps Mpps Layer 2 Aggregation 1 Line Card 10 L3 Uplink Gbps Gbps Layer 2 Aggregation 1 Line Card 11 L3 Uplink Mpps Mpps Layer 2 Aggregati
201. ut corrupting the model calculations and look ups This helps to make the model more robust and improves flexibility 14 2 2 Input sheets I Interface This sheet pulls in inputs from other parts of the model that are used by the Core model These are shown in tan colour coding together with a note indicating the original source of the input Access Co location or Consolidation This allows users to locate the original yellow or blue input cell which can be modified if required I Network_Elements This sheet defines all network elements used in the core model together with their relevant usage drivers I Product Demand This sheet contains the input cells for the traffic volume data from TDC for the following type of services Date 30 11 2009 Page 138 gt Telephony traditional and VoIP gt Broadband and Bitstream including Bitstream Transport IPTV and Video on Demand gt Access products PSTN ISDN raw copper gt Remaining non PSTN traffic such as leased lines mobile or cable TV Access products have been included to allow access specific costs in the core model to be allocated to the correct network increment The sheet also contains a section on xDSL which is used to cater for geotype targeting of active street cabinets and also the NGA scenario Call volumes for voice services are measured in call minutes and number of calls completed for each call type while volumes for data services are measured in subs
202. ve been taken into account in the revised Hybrid Model On the basis of this analysis NITA has updated the prices but kept the existing price trend of 3 for trench and for duct in both the primary and access networks 4 6 2 Copper cables The price of copper cables is to a large extent dependent on the price of raw copper Table 4 4 shows the figures for the price development of copper from Statistics Denmark for the period 2001 2008 The information indicates that the price of copper has been fluctuating significantly in recent years Table 4 4 Changes in the Wholesale Index for Copper Percent Pro Annum Cobber and Imported raw cobber articles thereof and semi manufactured 2001 2 1 6 3 2002 3 3 6 7 2003 3 1 5 2 2004 4 9 6 2 2005 6 0 8 9 2006 20 0 21 5 2007 9 8 8 3 2008 8 5 17 0 Average 2005 2008 6 8 5 4 Average 2001 2008 3 5 2 8 Source Statistics Denmark Information from TDC indicates a price trend of approximately 2 to 15 percent NITA has decided to leave the 6 price trend used in the previous model unchanged as this reflects the average price development of copper from 2005 2008 although again the prices themselves have been updated In 2005 NITA commissioned a study into the pricing of raw copper which supported NITA s view that prices for raw copper should be based on forward looking rather than historical costs Forward looking replacement costs will
203. xisting network structure As a consequence of this assumption the following types of exchange sites in TDC s network comprise the constraint in the Hybrid Model gt A Remote Subscriber Stage gt A local switch gt A transit or tandem switch Date 30 11 2009 Page 14 gt A DSLAM MSAN gt A site containing Layer 2 switching and or Layer 3 Routing equipment In Denmark a number of buildings are referred to as technical houses To the extent that these houses contain a concentrator and hence line cards they are accordingly deemed to fall within the definition of the scorched node 2 2 2 A Network Element approach The Hybrid Model adopts a network element approach to costing This means that the costs of network elements are estimated and the network elements bundled together to create the relevant interconnection products The actual bundling of the products will be determined by the routing factors for the particular services use of network elements 2 3 Services in the LRAIC Hybrid Model Telecommunications operators typically carry a wide range of services over their networks In addition to voice services operators provide leased lines broadband and other data services and other services such as cable TV The proportion of data related traffic has grown rapidly largely because of the growth in internet traffic This trend is likely to continue The models need to account for all of these services To exclude
204. zone can be turned into a circle through the use of a centralization factor This factor turns the exchange area into an Equivalent Homogenous Area EHA in which all subscribers are assumed to be evenly distributed The closer the centralization factor is to zero the closer the EHA is in terms of area to the original exchange area This is illustrated in the box below Date 30 11 2009 Page 66 Figure 8 2 Converting Exchange Areas into Equivalent homogenous areas Exchange Area A with centrality factor of 0 1 Exchange Area B with centrality factor of 0 9 Exchange Area A with a centralization factor of 0 1 Exchange Area B with a centralization factor of 0 9 Once the size of the zone has been estimated average beelines are calculated for different parts of the network namely gt ASM to PDP gt PDP to SDP gt SDP to NTP This is by far the most complicated part of the calculation algorithm and relies on the usage of a function that approximates the average distance between any point in the area and a particular point at a given distance from the centre of the area PDP areas are assumed to be circular and SDP areas are assumed to be rectangular but a fallback approach is in place for SDP areas so if the rectangular shape of the SDP area is not consistent with the circular shape of the PDP area the SDP area is considered to be circular as well The next step in the model is to convert beelines to realistic

Report on the LRAIC Model and User Guide

Contents

Download Pdf Manuals

Related Search

Related Contents