Relating IETF and TMF work Nigel Davis ndavis ciena

Relating IETF and TMF work Nigel Davis (ndavis@ciena. com) Mehmet Ersue (mehmet. ersue@nsn. com) Alex Zhdankin (Alex. Zhdankin@cisco. com) Jan Linblad (janl@tail-f. com) IETF #78 July 26, 2010 1

Overview • Aim – To identify key areas for collaboration between IETF and Tele. Management Forum (TMF) and to poll for interest and potential participants – To identify areas for potential work in IETF • Progression through the slides – Where is the area of focus for collaboration on work in IETF and TMF? – Why is this relevant? – What is taking place currently? – What are the next steps? 2

So why should you engage and how can you engage? • Consider as we progress through this material – Implications for operationalisation of industry convergence, value chains and cloud – Emergence of multi-layer, multi-vendor, multi-operator inter-working requires a converged network operations treatment across a very broad industry – The need to reduce the cost/complexity of operation and hence speed deployment aiming at an evolving 24 x 7 automated environment – Relevance of TM Forum in the broader industry including participation and activities – The common aims and goals for operations efficiency and automation • Working together in the converged industries – – – • Where can we work to operationalise the converged industries? What are the key areas for focus for common coherent development? How can your skills be applied to this broad challenge? How do the “data models” constructed for the network devices transform as they move through the OSS? How do we provide the information foundation for a more coherent and dynamic network operations environment? For more information see – www. tmforum. org for general information on TMF – http: //openoss. sourceforge. net/ and http: //sourceforge. net/apps/mediawiki/openoss/index. php? title=TIP_Main_Page for information on JOSIF (Joint Open Source Interface Framework) – http: //www. tmforum. org/community/groups/resource_management/default. aspx for information on Resource Management in TMF 3

Model to Implementation in TMF • Model Language and structure – UML (Unified Modelling Language) – Divided into many small component packages called ABEs (Abstract business Entities) • Tooling – – • Information model: RSM (Rational Software Modeller) from IBM Implementation: Tigerstripe (Open Source tooling via Eclipse {link}) from Cisco Version control: Subversion (via community) etc Model aspects – Abstract forms • Patterns and Architectures – Purpose oriented forms • Refactored encapsulations targeting specific needs – Implementation oriented forms • Refactored encapsulations targeting implementation – Specific implementations • • Some hand coded and some coded via Tigerstripe with reference implementations and test kits under sourceforge. net {add link} Model decoration via Stereotypes – Stereotypes used to drive interface implementation tooling – Lifecycle Stereotypes to guide application and use of the model • Challenges – Interrelating the multiple viewpoints – Dealing with versioning and migration – Intentionally refining the interrelationship between models and advancing beyond RFC 3444 4

Network models and interfaces in TMF • Key focus for this discussion – a focus on value for combined work – Management models for Network devices and network models – In the Information Framework (aka SID) this is the Resource ABE (Abstract Business Entity) • The SID is both Information Model and Data Model – Interface implementation forms in the Interface Program (aka TIP) • Resource model – Divides the problem into Logical (functional) and Physical • Logical Resource model – The key focus is on network functions – The SID includes both abstract and interface application specific models • Current position – Many deployments of the TMF MTNM (Multi-Technology Network Management) interface, a CORBA implementation – Some deployments with the newer TMF MTOSI (Multi-Technology Operation System Interface), a SOA implementation • Some of the ongoing work – Development of information patterns and architecture to aid mappings and evolution – Convergence of Connection Oriented and Connectionless models – Implementation forms generated complying with TIP Framework using JOSIF (Joint Open Source Interface Framework) tooling – Development of Business Services (Contracts and Tasks) and Order/Task grammar for task oriented interfaces 5

Examining some work so far… Focus here on: • Models – the value Other potential work focuses… not covered in detail here: • Interface infrastructure – the integration vehicle • Tooling and techniques – the enablers 6

Logical Resource ABE Model mapping SID • • The TM Forum interfaces utilise a Logical Resource model that can be applied at the nodal level Implementations of this model are widely deployed A valuable activity to reduce cost of integration between OSS and Network would be to “standardise” the mapping From an initial sketch it appears that: – We can map each If (and If. X) entry to a layer of the Termination. Point. – The If. Stack entries map to the association TPBound. To. TP. – The ent. Alias. Mapping. Table maps to the Physical Port name of the PTP potentially • • Work is ongoing to further strengthen the Logical Rsource model Some initial mapping work between MIBs and the Logical Resource model has been carried out in that context – Available to TM Forum Members and could be liaised in full or in part if 7 appropriate

More on models • Shows the logical resource model overlaid on a network device • This diagram is taken from SD 1 -18 (a supporting document of the MTNM/MTOSI model included in the download of TMF 814) 8

Open Source tooling potential target Common problem Concepts IETF Terms Common Terms TMF Terms Model in Yang Model in ECORE Model in UML Tool Open source tooling (e. g. JOSIF Tigerstripe) IETF Netconf Payload Framework Model translator: • Directive interpretation • Model pruning • Model transformations • Operation construction • XSD/WSDL and Java generation TMF SOA (TIP/JOSIF) Payload Framework Challenge mapping Yang models to UML models and vice-versa. 9

The value of engagement 10

Concept area overlap – speculation on insights from convergence Community DMTF IETF TMF TMF IETF “ 1990” 2010 2030 Concept convergence leads to terminology convergence and the emergence of a single community of focus groups the work of each intertwine to provide a set of consistent solutions for various specific cases. 11

So why engage? • Overlap of work in different bodies has reached the point where we must talk coherently – We are all essentially in the same Domain. – Now is the time to make moves towards language and model convergence – Opportunity to work in open source and to blend models from other SDO • Models do not need to be scary – sensible formality improves communication – – • Modelling more formally will enable IETF work to be used in other SDOs and vice-versa Clearer communication reduces cost/complexity and improves pace of deployment Inheritance, complex type, strong typing etc do not need to be used day one Need to converge languages and model translation to optimize model reuse Yang demonstrates recognised need to use model driven approaches in IETF – Increased awareness of benefits of top-down modelling to ease mappings and strengthen the solution – Yang tools for validation/translation available – Integration of YANG into Eclipse framework would enable progression to common framework • • Note: In general need to advance the tool chain in open environment (Eclipse/Source. Forge etc) Broaden the visibility and interpretability of your innovations – Models provide a way of expressing your work more clearly so it can be truly appreciated – Your work becomes more relevant/applicable to other SDOs in a broader context – Working within a broad industry context and consistent language will strengthen your work 12

Conclusions • Aim – To identify a key area for collaboration between IETF and TM Forum and to poll for potential participants – To identify areas for potential work in IETF • Progression through the slides – Where is the area of focus for collaboration between IETF and TMF • • Common aims and goals Network Resource model Use of common open source tooling Potentially in future… Interface implementation convergence – Why is this relevant? • • • Working together towards the future industry convergence Working to reduce the cost/complexity of operation and hence speed deployment 24 x 7 automation across cloud and B 2 B value chains etc – What is taking place currently? • Logical Resource work related to some Yang modules including some MIB mappings – What are the next steps? • Working together in the converged industries – Where can we work to operationalise the converged industries? • TM Forum is currently at the axis of operationalisation of networks – What are the key areas for focus for common coherent development? • • • Resource model Modelling and mapping of models – Business case Infrastructure and network convergence – How can your skills be applied to this broad challenge? • we need more IETF engagement and MIB expertise 13

How can you engage? – Next steps • Possible joint work we could address: – Understand the relation between TMF Resource model and future YANG modules • How to prepare future YANG modules for ad-hoc adoption by other SDOs? – Potentially aiming for an aligned network information model? – Deriving YANG models from TMF Resource model (see also ongoing work: draftlinowski-netmod-yang-abstract-03. txt) – Prepare MIB mappings to SID Resource model or review MIB mappings prepared at TMF – Identify a converged target model for connectionless connectionoriented networks • Possible mode of cooperation: – Provide a bridge between TMF and IETF and join activities under ITU-T JCA-mgt umbrella • TMF Logical Resource Alignment activity or other model and methodology harmonization activities • Currently one person is actively supporting this – Review of TMF documents per liaison – Joint chartered work item in the O&M area workgroup with a concrete problem statement • Seems to be the best cooperation approach so far 14

Any questions? 15