Software Requirements Specification
Contents
1. int amp tag const char NULL const char abstract gt setQualityBad void abstract gt inseri value const std string amp tag const char NULL void abstract gt setQuaityBad reason const char void abstract gt insert value const char tag const char NULL void abstract setQualityJ ncertain void apstract insert vdue DipBool tag const char NULL void abstract gt setQualityJ ncertain reason const char void abstract gt insert value DipByte tag const char NULL void apstract insert vdue DipShort tag const char NULL void abstract gt insert value Dipint tag const char NULL void abstract insert vdue DipLong tag const char NULL void lt lt abstract gt gt insert vaue DipFloat tag const char NULL void lt lt abstract gt gt insert vdue DipDouble tag const char NULL void nages lt lt abstract gt gt insert addr const DipBool size int tag const char NULL void E abstract insert addr const DipByte size irt tag const char NULL void nbnages H abstract gt insert addr const DipShort size int tag const char NULL void abstract gt insert addr const Dipint size irt tag const char NULL void abstract gt insert addr cons2. distribution mechanism DIP doesn t support the transport of services e When there is a need to carry high level information between large heterogeneous systems E g between several controls systems e When this information is valuable to multiple systems at once This is also a reason why DIP shall be used to carry high level information rather than dedicated information It has also an influence on the structuring of the data propagated through DIP DIP is particularly designed to carry live data or near real time data e When the data conveyed isn t critical 1 e it doesn t imply security or safety issues Beside the plain DIP libraries that are freely available for Windows and Linux platforms with C and Java APIs there are DIP plug in components being developed by CERN for both LabView and WINCC OA PVSS 1 7 Basic DIP principles In one sentence for the techies a multicast data distribution mechanism using the push model DIP is essentially an information distribution service and as such may be compared with subscribing to news papers or magazines If one wants to receive a magazine he or she may subscribe to that magazine by giving the name of the magazine to the supplier Whenever a new edition of that magazine is published that person will receive a copy of it The person is of course not restricted to subscribing to one magazine or just to magazines from a single publisher The person need only provide to the sup
3. only to a subset of its content e There is no filtering mechanism provided by DIP itself 1 e one subscribed to a Publication it will receive all its updates The Publisher has full control over data content quality and timestamp There is a notion of data contract between the Publishers and their Subscribers A Publisher responsible for a set of Publications is not allowed to change online the structure of its Publications once they have been made available for potential Subscribers e There are no particular security mechanisms implemented in the DIP protocol For example a Subscriber has no means to authenticate the source of the data it receives Similarly the DIP protocol doesn t offer a Publisher the possibility to restrict access to a set of authenticated Subscribers Another example the Publications names are not nominative In other words when a Publisher stops the Publications names it was using are from then on freely available to other Publishers e There is no feedback to a Publisher that the data published actually reached its Subscribers also known as one way communication Hence there is no retransmission in case of a data delivery issue Public CERN 2011 12 19 Page 7 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc 1 9 DIP dynamics The pseudo UML schema below describes the sequence of actions between a Subscriber and a Publisher Subscriber DNS Pub
4. size irt amp tag const char NULL const DipByte abstract send value const DipBool size int timestamp const DipTimestamp amp void const abstract gt extractshort amp rray size int amp tag const char NULL const DipShort abstracb gt send value const DipByte size int timestamp const DipTimestamp amp voi d const apistract extract nt amp rraysize int amp tag const char NULL const Dipint abstract gt send value const DipShort size int timestamp const DipTimestamp amp void const abstract gt exractFloatArrayize irt amp tag const char NULL const DipFlaat abstract send value constDiplnt size int timestamp const DipTimestamp amp void constabstract exractDoubleArraysize int amp tag const char NULL const DipD ouble apstract gt send value const DipLong size int timestamp const DipTimestamp amp void lt lt const Abstract gt exractLongArray size irt amp tag const char NULL const DipLong abstract gt send value const DipFloat size int timestamp const DipTimestamp amp void const abstract gt extractStringArray size int amp tag const char NULL const std string apstract send value const DipD ouble size irt timestamp constDipTimestamp amp void 9 corist abstract gt exractCStringArray size
5. 4 Publishers shall provide a quality reason for changing their Publications quality to Bad or Uncertain Finally the timestamp provided with each Publication s update is meant to describe when the data provided was actually measured This could be different than the time at which this update is sent to potential Subscribers Recommendation 25 Publishers shall provide a correct timestamp for each of their Publications update END OF DOCUMENT Public CERN 2011 12 19 Page 20
6. 9 Apply a consistent logic to structure the Publications offered by the Publishers 12 Recommendation 10 Apply a consistent logic in the Publications intrinsic characteristics names internal structure and refresh rates 12 Recommendation 11 Adjust the granularity of the Publications content for their efficient usage on the Subscribers side 12 Recommendation 12 Recommendation 13 Publications names Recommendation 14 Recommendation 15 Recommendation 16 times Recommendation 17 Recommendation 18 Recommendation 19 issues Recommendation 20 Recommendation 21 interface s methods Recommendation 22 DipPublicationErrorHandler callback Recommendation 23 Recommendation 24 Uncertain Recommendation 25 Public For uniformity preferably use American English for publications names and descriptions 13 Make use of standard DIP characters and for uniformity prefer the delimiter for the 13 Follow the DIP Publications naming convention described in Ref 1 13 Describe accurately the DIP Publications static and dynamic characteristics 13 Make sure the DIP Publications characteristics are maintained up to date and available at all 13 Communicate any foreseen changes on the DIP Publications for smooth transitions 13 Private DNS s shall not be visible outside their hosting domain 14 Make use of the provided DIP test tools and Browser to detect and diagnose communication 14 Use unique names
7. CERN DIP usage recommendations EN ICE Version 1 1 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc Revision History 7 March 2008 Implemented Wayne Salter comments Mathias Dutour 25 April 2008 Implemented Renaud Barillere comments Mathias Dutour 14 December 2011 Fixed various code samples updated web Brice COpy links Public CERN 2011 12 19 Page 2 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc Table of Contents 1 Introduction 5 1 1 Purpose 5 1 2 Intended Audience 5 1 3 Acronyms and definitions 5 1 3 1 Acronyms 5 1 3 2 DIP definitions 5 1 4 References 5 2 Understanding DIP basics 6 2 1 Information over DIP 6 2 2 DIP applicability 6 2 3 Basic DIP principles 6 2 4 Specific DIP characteristics 7 2 5 DIP dynamics 8 3 Managing DIP deployment and communication 9 3 Cross domains DIP communication on the TN 9 3 2 Special case WinCC OA PVSS as a filtering mechanism 10 3 3 Additional recommendations on Publishers Subscribers Publications and DNS 12 3 3 1 Publishers and Subscribers 12 3 3 2 Publications 12 3 3 3 DNS 14 3 4 Technical information concerning deployment 14 4 Using the DIP API 15 4 1 The big picture 15 4 2 DIP Factory 16 4 3 Subscribers 17 4 4 X Publishers 18 4 5 Publications 19 Table of Figures Figure 1 DIP dynamics 8 Figure 3 Cross domain DIP communication on th
8. Definition 2004 Wayne Salter on behalf of the LDIWG EDMS Reference 457113 Ref 2 DIP Name servers Principles and mechanics 2006 Mathias Dutour Ref 3 DIM http cern ch dim Clara Gaspar Public CERN 2011 12 19 Page 5 User manual DIP usage recommendation Date 14 December 201 1 DIPUsageRecommendations doc Understanding DIP basics 1 5 Information over DIP Besides this document and recommendations there are other sources of relevant information addressing various DIP viewpoints One could mention e The DIP web pages that presents the DIP news and the available software to download http cern ch dip e For any DIP questions one could contact DIP experts via our the support email icecontrols support cern ch e The EN ICE Piquet is available 24 7 to investigate DNS related issues outside working hours on CERN number 164930 1 6 DIP applicability As mentioned in Ref 1 DIP can be described as DIP is a system which allows relatively small amounts of real time data to be exchanged between very loosely coupled heterogeneous systems These systems do not need very low latency The data is assumed to be mostly summarized data rather than low level parameters from the individual systems i e cooling plant status rather than the opening level of a particular valve In particular DIP is a recommended candidate to deal with the following situations e Where there is a need for a simple integration of a data
9. EPT Finally the DIP delivery is provided with sample programs which are very useful to check your DIP environment health and to use as a starting point for real development Additionally the DIP Browser is also an important tool that can connect to any DNS given that the visibility is granted and the DIPNS environment variables properly set It is therefore often useful to use the DIP Browser to detect and diagnose DIP communications issues Recommendation 19 Make use of the provided DIP test tools and Browser to detect and diagnose communication issues Public CERN 2011 12 19 Page 14 User manual DIP usage recommendation Date 14 December 201 1 DIPUsageRecommendations doc 3 Using the DIP API 3 1 The big picture The DIP delivery contains mainly 3 sets e The DIP libraries C and Java e The DIP Browser a utility tool that connects to the DNS that actually contains a Publishers that provides the list of Publications it manages e Sample C and Java programs to give DIP a spin to present concrete examples and to have a starting point for real development Although the packages interfaces and classes presented in this document originate from the C library the most commonly used so far all the recommendations apply as well for the Java library Here the DIP packages with their interfaces and dependencies are presented e from T tools from i Interf
10. P should not be used as a mechanism to send critical status or operational commands to your systems 2 3 2 Publications One could apply the previous recommendations to the Publications themselves Hence it is important to properly apply a consistent logic across your Publication when you define and structure them This logic shall be natural as it defines the way the Subscribers are going to access this data Often this Publication s organization is driven by the logical structure of a system which shall makes sense for the Subscribers and less often by the hardware architecture or other considerations Finally adjusting the granularity of the Publication s content will help to maximize usability and extensibility avoiding the super Publication with dozens of data fields just like the atomic one containing little valuable information by itself Recommendation 9 Apply a consistent logic to structure the Publications offered by the Publishers Recommendation 10 Apply a consistent logic in the Publications intrinsic characteristics names internal structure and refresh rates Recommendation 11 Adjust the granularity of the Publications content for their efficient usage on the Subscribers side Public CERN 2011 12 19 Page 12 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc Each Publication s name must be unique per DNS instance hence a naming convention was defined in Ref 1 and sum
11. S one has to use the DNS provided in the DIM delivery see Ref 3 One shall make sure it is deployed on a machine only visible within the private domain to prevent any risks of duplications and confusion concerning cross domain Publications managed by the centrally managed DIP name servers Recommendation 18 Private DNS s shall not be visible outside their hosting domain 2 4 Technical information concerning deployment Deploying DIP applications is rather simple one has to make sure that e Either the centrally managed DIP name server is visible to Publishers and Subscribers in which case there are no particular settings to be adjusted e Or that the domain private DIP name server or servers are visible In the second situation the DNS DIM DNS PORT environment variable has to be set to 2506 This is the port used by DIP to accept incoming connections On the Publishers and Subscribers side the environment variable DIPNS shall point to the name of the machine where the private DNS is running The value associated to DIPNS can be a comma separated list of machine names hence allowing failover DNS to be declared e g DIPNS pcitcol 61 cern ch pcitcol62 cern ch Finally the DIP Publishers are using the ports in the range 5100 to 6000 so these shall be opened on their hosting machines e g on a Linux machine iptables A RH Firewall 1 INPUT m state state NEW m tcp p tcp dport 5100 6000 j ACC
12. ace Suppor tools Data Exceptions Interface e e DipSubscription DipSubscriptionListener PHBIUDIIESUOR from Subscriber Interfaces from Subscriber Interfaces hd irom Publisher Interiaces N 7 DipPublicationErrorHandler DipUpdateRequestListener from Publisher Interfaces from Publisher Interfaces Subscriber Interfaces z Publisher e Interfaces DipFactory from DIP Core DIP Core Figure 4 DIP packages interfaces Exceptions Predefined Data types ARR uote AAR Publisher Data Subscriber Interfaces gt Interface Interfaces Z A DIP Core Figure 5 DIP packages dependencies Public CERN 2011 12 19 Page 15 DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc All UML diagrams of DIP are available in Error Reference source not found Each aspect is refined in the following sections 3 2 DIP Factory Dip from DIP Core theF actory Dipfother const Dip amp gt operator other const Dip amp Dip amp lt lt static gt gt create name const char DipFactory DipFact S static create DipFactory Ipractory static shutdown void from DIP i lt lt irtual gt gt DipFactory abstract createDipSubscr
13. act gt setQuality Uncertainfreason const char void Figure 8 Publishers involved classes The Publishers must provide an instance of the DipPublicationErrorHandler class and properly handle the potential error messages received Recommendation 22 Publishers process adequately the messages received through the DipPublicationErrorHandler callback Public CERN 2011 12 19 Page 18 DIP usage recommendation Date 14 December 201 1 DIPUsageRecommendations doc 3 5 Publications Here the main classes relevant to the Publications are presented manages 4 DipPublication DipData from Publisher Interfaces from Data Interface v DipPublicati on amp DipD ta const DipData amp 9 virtual gt DipPublication operator cther const DipData amp DipD ata amp const abstract gt getTopicName const char DipData abstract send value const DipD amp ta amp timestamp const DipTimestamp amp vaid lt lt Virtual gt gt DipData abstract send value const char timestamp const DipTimestamp amp void const abstract gt exraftstring tag const char NULL const std string amp apbstract send value const std string amp timestamp const DipTimestamp amp voi d constabstract gtractC String tag const char NULL const char a
14. amp void abstract send value DiplInt timestamp const DipTimestamp amp void lt lt abstract gt gt send value DipLong timestamp const DipTimestamp amp void lt abstract gt send value DipFloat timestamp const DipTimestamp amp void abstract gt sendivalue DipDouble timestamp const DipTimestamp amp void abstracts send value const char size int timestamp const DipTimestamp amp void abstracts send value const std string size int timestamp const DipTimestamp amp void abstracts send value const DipBool size int timestamp const DipTimestamp amp void abstract send value const DipByte size int timestamp const DipTimestamp amp void abstract send value const DipShort size int timestamp const DipTimestamp amp void S zabstract send value const Diplnt size int timestamp const DipTimestamp amp void abstract send value const DipLong size int timestamp const DipTimestamp amp void apstract send value const DipF loat size int timestamp const DipTimestamp amp void abstract send value const DipDouble size int timestamp const DipTimestamp amp void abstract setQuality Bad void abstract setQuality Bad reason const char void lt lt abstract gt gt setQuality Uncertaino void lt lt abstr
15. ant for the Publishers to assign timestamp quality and quality reason to DIP Publications correctly and for the Subscribers to understand them Although the Subscribers could receive 4 different quality levels a Publisher can either set the quality of a Publication to Bad or Uncertain given that normal updating assigns a good quality and Uninitialised is managed by DIP Quality Description DIP QUALITY UNINITIALIZED Indicates the DIP data object contains uninitialised data DIP QUALITY BAD Indicates that data can not be used DIP QUALITY GOOD Indicates that data can definitely be trusted The last update from the publisher was successful and the value s that this quality corresponds to are the most current available DIP QUALITY UNCERTAIN Indicates that the last update attempt by the publisher had failed meaning the Publication data source was not accessible The value s this quality corresponds to is the last known good value by the Publisher which can no longer be considered up to date Public CERN 2011 12 19 Page 19 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc Recommendation 23 Make proper use of the DIP quality levels on Publisher and Subscribers side Additionally the Publisher has the possibility via the API to provide for Bad and Uncertain qualities a reason for setting these quality levels always useful to the Subscribers Recommendation 2
16. base is used as a bridge here with a Publisher connected to the central DNS another connected to the private DNS One could use such mechanism when only a subset of the total data available inside the domain shall be exposed across domains possibly structured differently into Publications Public CERN 2011 12 19 Page 10 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc There are several advantages to this technique e Itis simple to setup in a WINCC OA PVSS project only using DIP v 5 2 0 and above e tis a flexible mechanism to adjust visibility of Publications And disadvantages e There is a maintenance effort required to guaranty the availability of the private DNS in the domain e It could become difficult to maintain a clear picture of the Publications visibility in case there are many Publishers using this bridge mechanism in WINCC OA PVSS Public CERN 2011 12 19 Page 11 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc 2 3 Additional recommendations on Publishers Subscribers Publications and DNS 2 3 1 Publishers and Subscribers In order to keep the big picture of your DIP data flows clear and manageable in a domain it 1s important not to distribute the DIP data over too many or two few Publishers and Subscribers This is achieved by clearly identifying high level responsibilities and then mapping the latter onto Publ
17. bstract send value DipBod timestamp const DipTimestamp amp void constabstract etractBool tag const char NULL DipBool abstract send value Di pByte timestamp const DipTimestamp amp void const abstract gt eractByte tag const char NULL Dippyte abstract gt send value DipShort timestamp const DipTimestamp amp void const abstr ct gt exractshort tag constchar NULL DipShort abstract gt send value Diplnt timestamp const DipTimestamp amp void const abstract extractint tag constchar NULL Diplnt abstract gt send value DipLong timestamp const DipTimestamp amp void const abstract gt exractFloat tag const char NULL DipFlaat abstract gt send value DipFloat timestamp const DipTimestamp amp void lt lt const ab tract gt exractDouble tag const char NULL DipDouble abstract gt send value DipD ouble timestamp const DipTimestamp amp void const abstract gt exractLong tag const char NULL DipLong abstract gt send value const char size int timestamp const DipTimestamp amp void const abstract exractBoolArray ize int amp tag const char NULL const DipBool abstract send value const std string size int timestamp const DipTimestamp amp void const abstragt gt exractByteArray
18. e TN 9 Figure 4 Special case WinCC OA as a filtering mechanism 10 Figure 5 Error Bookmark not defined Figure 6 DIP packages interfaces 15 Figure 7 DIP packages dependencies 15 Figure 8 DIP Factory involved classes 16 Figure 9 Subscribers involved classes 17 Figure 10 Publishers involved classes 18 Figure 11 DIP Publications main classes 19 Public CERN 2011 12 19 Page 3 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc Table of Recommendations Recommendation 1 Use the central DNS for any cross domain DIP data exchanges 9 Recommendation 2 Use the central DNS whenever possible including for intra domain DIP data exchange 9 Recommendation 3 Identify your domain private and cross domains DIP data 9 Recommendation 4 As a Publisher of cross domain data maintain up to date the information over your Publications available for Subscribers 10 Recommendation 5 Do not regroup Publishers for internal domain data with Publishers for cross domain on the same machine s to simplify communication management 10 Recommendation 6 Limit the Number of Publishers and Subscribers per domain to achieve good DIP manageability 12 Recommendation 7 Make sure each Publisher and Subscriber is assigned only one clear high level responsibility 12 Recommendation 8 DIP should not be used as a mechanism to send critical status or operational commands to your systems 12 Recommendation
19. h as when Publications are known not to be available maintenance period etc Recommendation 15 Describe accurately the DIP Publications static and dynamic characteristics Of course this information is primarily addressed to the Subscribers but also relevant to maintain consistency on the Publisher s side for example when extensions have to be implemented e g Integrating new equipment leading to new Publications So it shall be maintained up to date by each domain and available at all time preferably in one place only E g web page interface definition document IDD etc It is equally important to communicate any changes foreseen that could affect the Publications in place to guarantee smooth transitions and avoid clashes Recommendation 16 Make sure the DIP Publications characteristics are maintained up to date and available at all times Recommendation 17 Communicate any foreseen changes on the DIP Publications for smooth transitions Finally occasionally one asks whether there should be dedicated Publications playing the role of watchdog or heartbeat to show the health of Publishers in WINCC OA PVSS The point is often raised when dealing with very slow updating Publications raising some doubts on the Subscribers side whether the Publisher 1s alive or not Although there isn t any constraint there this is in principle not necessary for 2 reasons e First each Publication has a quality that indicates whether
20. his picture represent the future DIP organization to be deployed mid 2008 with a central DNS on the TN and various Publishers and Subscribers communicating together across domains This DNS is maintained by ITCO and therefore will require no maintenance for the experiments Recommendation 1 Use the central DNS for any cross domain DIP data exchanges Recommendation 2 Use the central DNS whenever possible including for intra domain DIP data exchange Within a certain domain the DIP communication 1s typically not constrained and DIP data can be freely exchanged between Publishers and Subscribers In case there 1s a need to restrict some DIP data access within a certain domain to a set of Subscribers this can be achieved by network configuration filtering on IP addresses However this filtering is not so flexible and supposes an all or nothing filtering logic on the Publications Then the picture above clearly indicates the connections across domains must be clearly identified and managed This 1s achieved by different means e First it is important to identify properly which DIP data is of interest to other domains to prevent to mix domain private and cross domain information within the same Publications Recommendation 3 Identify your domain private and cross domains DIP data Public CERN 2011 12 19 Page 9 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc e tis crucial for the Publishers
21. icular the connected disconnected and handleException methods shall not be neglected as they indicated important information about the health and the status of the connection to the user application s Publishers Recommendation 21 Subscribers shall implement and process adequately the entire DipSubscriptionListener interface s methods Public CERN 2011 12 19 Page 17 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc 3 4 Publishers Here the main classes relevant to the Subscribers are presented DipPublicationErrorHandler DipPublication from Publisher Interfaces from Publisher Interfac es DipP ublicationQ abstracts handleExc eption publication DipP ublication ex DipException amp void lt virtual DipPublic ationd lt lt const abstract gt gt getTopic Named const chart lt lt abstract gt gt send value const DipData amp timestamp const DipTimestamp amp void lt lt abstract gt gt send value const char timestamp const DipTimestamp amp void P amp abstracts send value const std string amp timestamp const DipTimestamp amp void implem ERE ETC S abstract send value DipBool timestamp const DipTimestamp amp void amp abstracts send value DipByte timestamp const DipTimestamp amp void lt lt abstract gt send value DipShort timestamp const DipTimestamp
22. ing Interface DCS Detector Control System DIP Data Interchange Protocol DIM Distributed Information Management DNS DIP Name Server DSS Detector Safety System EN Experimental Network GPN General Purpose Network EN ICE EN Department Industrial Controls Group PVSS ProzessVisualisierungs und SteuerungsSystem WINCC OA WinCC Open Architecture previously known as PVSS SCADA Supervisory Control And Data Acquisition SLS Service Level Status TN Technical Network UML Unified Modeling Language 1 3 2 DIP definitions For easy identification every occurrence of a name depicted in these definitions will start with an uppercase letter in the rest of the document Publisher A producer of DIP data independently of software program and applications considerations A Publisher is responsible for the definition of the structure the content and the provision of the DIP data to its Subscribers Subscriber A client of DIP data independently of software program and applications considerations Publication A logical container that structures related atomic pieces of DIP data This logic container has a name as well as timestamp and quality properties that describe its state This Publication gets its content updated at will by its producer Subscription A Subscription refers to a Subscriber s interest in receiving updates of a certain Publication 1 4 References Ref 1 LHC Data Interchange Protocol DIP
23. iption topicname const chart listener DipSubscriptionListener Dipsubscription abstract createDipP ublic ation topicname const char errHandler DipPublicationE morHandler DipP ublic ati on abstract destrayDipSubscription s ubscription DipSubscription void abstract destroyDipPublication publication DipPublication void abstract createDipData DipData abstract createDipBrowser DipBrows er v9 DipFactory Figure 6 DIP Factory involved classes The Dip class is an entry point for any Publisher or Subscriber to access the Singleton DipFactory class One important point it is recommended to use the characterized create method of the Dip class passing a unique name as this name will be used to identify the user application handle in the DNS Hence one could use such construction to guarantee the name is unique when running the user program DIP object DipFactory dip DipTimestamp aTimeStamp long someRandomness long aTimeStamp getAsMillis string clientName MyClientName clientName ltoa someRandomness dip e DripircreavtefolientName c SUEQOS Then this DipFactory class allows the user code to manage the life cycle management of Publications and Subscriptions Recommendation 20 Use unique names when spawning the DipFactory singleton Public CERN 2011 12 19 Page 16 DIP usage recommendation Date 14 Decembe
24. ishers and Subscribers Often a Publisher is responsible for a horizontal technical service such as Cooling and Ventilation DSS etc and therefore 1s in charge of an homogeneous and logically related set of information On the other hand a Subscriber is rather in charge of a vertical operational slice such as an experiment s DCS relying on many technical services So again these responsibilities shall be well scoped to balance flexibility versus complexity Recommendation 6 Limit the Number of Publishers and Subscribers per domain to achieve good DIP manageability It is in general advisable to assign each Publisher and Subscriber only one high level responsibility for example to prevent the unavailability of unrelated Publications in the event of maintenance Recommendation 7 Make sure each Publisher and Subscriber is assigned only one clear high level responsibility Last but not least DIP 1s meant to carry non sensitive non critical data due to the minimum security and data delivery it offers and also due to the lack of visibility on the origin of a Publication Be aware that when using DIP to send a command there is no feedback telling whether the command was received interpreted or ignored See 1 8 Specific DIP characteristics Hence DIP is either not meant to carry mission critical commands like Turn on system X but may be used with caution for supporting tasks like Activate logging Recommendation 8 DI
25. lisher 1 Subscribe to D baad 2 Waits unil dip PlantStatus T CE available 3 Publish dip PlantStatus 4 Provide Server handle 5 Process data ee r e P4 6 Publish dip PlantStatus update Figure 1 DIP dynamics The picture above provides a simplified picture of what s happening when one subscribes to some Publications here dip PlantStatus as well as the role of the DNS in connecting the Publishers with their Subscribers More detailed explanations can be found in the DIP FAQ http cern ch en ice DIP FAQ From this picture it is easy to understand that the network settings shall be configured adequately to allow the correct communication These aspects are covered in the next chapter Public CERN 2011 12 19 Page 8 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc 2 Managing DIP deployment and communication 2 1 Cross domains DIP communication on the TN The picture below shows the logical data flows between DNS Publishers and Subscribers on the TN and on the ENs Permitting communication tr tha TNI to the TN Receiv a 2 Dedicated network E g LHC um experimental network as E Subscriber Other dedicated network Figure 2 Cross domain DIP communication on the TN T
26. marized below dip Domain SubDomain SubSubDomain 1 SubSubDomainx Item E g dip ATLAS Muon EndCap Rack Rack53 Temperature Finally please use the standard characters abcdefghijkImnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXY7Z01234506789 and delimiters for the Publications names as one couldn t foresee their Subscribers usage Recommendation 12 For uniformity preferably use American English for publications names and descriptions Recommendation 13 Make use of standard DIP characters and for uniformity prefer the delimiter for the Publications names Recommendation 14 Follow the DIP Publications naming convention described in Ref 1 Then it is important to communicate efficiently DIP detailed information to respect the DIP data contract between the Publishers and the Subscribers Starting from the organization of the Publications themselves then other Publication s level details shall be provided e The description of the Publication s purpose e The description of every field of the Publication together with the unit intended for each field e How the Publication is refreshed cyclically periodically manually on a particular event etc and possibly how often e How the timestamp of the Publication is affected by the Publisher e g when reading the data from equipment arbitrarily etc How the DIP quality of the data is assigned to the Publication e Any other useful information suc
27. one can trust or not the content of the Publication e And second the connection s status with a Publisher is known via some dedicated callbacks on the Subscriber s side Now there is a situation where a watchdog could be useful One could for instance use a Publication as a watchdog when the data produced by a Publisher comes from another system e g a physical device which could be blocked Public CERN 2011 12 19 Page 13 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc itself and therefore not updating its own data anymore In this case a watchdog could be implemented between the Publisher and this system and then published via DIP for the Subscribers Le to assert the health of the device not of the Publisher 2 3 3 DNS The role of the DNS 1s to maintain the list of Publications available and connect the Subscribers to the Publishers It is therefore necessary at the time this connection must be established The hot DNS managed by ITCO is backed up by a standby one in a failover fashion and the guarantee is granted that the switch is transparent for the Publishers and Subscribers More details of this mechanism are provided in Ref 2 Although there aren t official scalability performance figures at this time no particular issues have been noticed for example when switching more than 20k Publications from the hot to the standby DNS To setup a private DN
28. plier names of the publications he or she is interested in they need not know who or where the publisher actually is and they will receive new editions of the requested publication as they become available subscription DIP is essentially playing the role of the Supplier in the above scenario There is one notable difference between DIP and the magazine scenario described above That is a magazine is published on a periodic basis this is not necessarily the case with DIP publications which are published at a time decided by the publisher Public CERN 2011 12 19 Page 6 User manual DIP usage recommendation Date 14 December 2011 DIPUsageRecommendations doc Important components in the DIP architecture are Publishers Subscribers and Publications as shown in the diagram below DIP Client DIP Server Subscribe Publishe See the chapter 3 Using the DIP API for more details about Publishers Subscriber and Publications from a DIP API point of view 1 8 Specific DIP characteristics There are some important characteristics about DIP one should be aware of e The Publisher and its Subscribers don t know each other explicitly The DIP protocol actually the DNS connects the Subscribers with their Publishers of interest A subscribers knows at all time the status of its connections to Publishers e A Publication is a structured container of atomic data It is a consistent item and the Subscribers cannot subscribe
29. r 201 1 DIPUsageRecommendations doc 3 3 Subscribers Here the main classes relevant to the Subscribers are presented DipDIMInternalError from Exceptions l DiplnternalError ae DipDIMInternalError DiplntemalError DipDIMInternalError DiplntemalError S getErrorCode TypeMismatch from Exceptions ee STweMismatch TypeMismatch TypeMsmatch DipException from Exceptions p DipException vittual DipException V L BadParameter from Exceptions Haz y c mmi cac ae fhrows BadParametert BadParameter BadParameter HERR uses DipSubscription DipSubscriptionListener from Subscriber Interfaces from Subscriber Interfaces lt lt abstract gt gt getTopicName const char virtual DipSubscriptionListener cabstract requestUpdate void S abstract handle Message subscription DipSubscription message DipData amp void F Dip Subscription lt lt abstract gt gt disconnected subscription DipSubscription reason char void S virtual Dip Subscription cabstract connected subscription DipSubscription void abstract handleException subscription DipSubscription ex DipException amp void implements Figure 7 Subscribers involved classes The user application implementing the DipSubscriptionListener must implement all its methods In part
30. t DipLong size irt tag const char NULL void 1 abstract gt insert addr const DipFloat size irt tag const char NULL void l abstract gt insert addr const DipD ouble size int tag const char NULL void abstract gt insert value const std string size int tag const char NULL void abstract gt insert value const char J size int tag const char NULL void 1 from Data Interface lt lt const abstract gt gt ge vaueType tag const char NULL DipDataType 4 constabstract getTags nTags int amp constchar lt lt enum gt gt pipTimestamp constabstract contains tag const char bool DipQuality DipTimestamp const abstract gt size int from Data Interface setCurrertTime const abstract isEmpty bool eDIP QUALITY UNINITIALIZED 2 0 const abstract gt extractDataQuallity DipQuality eDIP QU ALITY BAD o set anos lt lt const abstract gt gt eractR easonF orQual itx const std string amp F F const g amp AsMillis c tabstract gt gt exrractDipTime const DipTimestamp amp DIP QUALITY GOOD 1 9 0 bis SPa p P DIP QUALITY UNCERTAIN 2 const geAsNanos lt lt const abstract gt getvalueDimension tag const char unsigned Figure 9 DIP Publications main classes As indicated earlier it is import
31. to present the data available for Subscribers and to maintain this information up to date Once a Subscriber is interested in DIP data being published in another domain its responsible could contact the Publisher s responsible and together decide to setup a proper network configuration enabling the DIP communication Recommendation 4 As a Publisher of cross domain data maintain up to date the information over your Publications available for Subscribers e Finally these Publishers providing data for other domains could be collocated on the same machine s to simplify communication management 1 e network configuration Furthermore Publishers of intra domain data shall preferably not be collocated with Publishers for cross domain data Recommendation 5 Do not regroup Publishers for internal domain data with Publishers for cross domain on the same machine s to simplify communication management 2 2 Special case WinCC OA PVSS as a filtering mechanism The picture below describes how one could use a WinCC OA system and a private DNS as a DIP data filter Technical Network Subscriber Dedicated network E g LHC experimental network Figure 3 Special case WinCC OA as a filtering mechanism In this configuration a domain private DNS is used to manage all the DIP data that shall remain within the domain while the central DNS is used to expose and exchange DIP data with other domains Then the WINCC OA PVSS internal data
32. when spawning the DipFactory singleton 16 Subscribers shall implement and process adequately the entire DipSubscriptionListener 17 Publishers process adequately the messages received through the 18 Make proper use of the DIP quality levels on Publisher and Subscribers side 20 Publishers shall provide a quality reason for changing their Publications quality to Bad or 20 Publishers shall provide a correct timestamp for each of their Publications update 20 CERN 2011 12 19 Page 4 User manual DIP usage recommendation Date 14 December 201 1 DIPUsageRecommendations doc 1 Introduction 1 1 Purpose This document comes in the context of the role of DIP for the LHC operations and is meant to help users to manage deploy and configure DIP to match their needs It addresses the current situation with the DNS located on the General Purpose Network GPN as well as when the DNS will be located on the Technical Network TN Several topics are covered here ranging from DIP basic understanding to management and organization of the DIP data distribution and DIP API usage High and low level recommendations are provided in various situations 1 2 Intended Audience This document is meant for any person dealing with data distribution via DIP It addresses the concerns of team managers network engineers developers and integrators 1 3 Acronyms and definitions 1 3 1 Acronyms API Application Programm
Download Pdf Manuals
Related Search