Skip to end of metadata
Go to start of metadata

Date

11 December 2017

Attendees

  • Lyndon Ong

  • @Tracy van Brakle

  • Nigel Davis
  • Kam Lam
  • Karthik Sethuraman
  • Italo Busi
  • Bernd Zeuner

Goals

Discussion Items

TimeItemWhoNotes
 
  1. TAPI 2.0 SDK Last Call
Lyndon
  • TAPI 2.0 Release Candidate 3 posted by Karthik
  • RC3 is the basis for a 2 week Last Call prior to posting TAPI v2.0.0 – Last Call ends December 20th, after which the OTCC TST will decide if there are any problems with approving the SDK
  • Approval of TAPI v2.0.0 will be very helpful to activities such as the OIF 2018 demonstration which have been waiting on completion of TAPI 2.0 specs – OIF people were very happy to hear about the Last Call
  • We are now on RC3 of TAPI 2.0 so most bugs/issues should already have been discussed and resolved
  • Comments should be made on the Snowmass github site, and focus of comments should be on bugs/issues in the Release Candidate and not on new functions or enhancements – these will be deferred to a future TAPI v2.0.x
  • Expectation is that we will identify further issues during the development of test specifications, implementation and interop testing of the OIF 2018 demonstration and that these will go into TAPI v2.0.1, v2.0.2, etc. as was the process with TAPI v1.0 and OIF 2016 testing.
  • Italo - some concern with potential for changes - should we continue with RCs instead? Backward compatibility a concern. Karthik - haven't seen any compatibility issues in the previous RCs, feels ready for LC - note that there are still discussion on the use cases but not affecting the model itself. No major change requests in last 6 months. Nigel - agrees that model has not been changing, have been looking at different use cases. Example - node can be modeled as single or multilayer using the model, but different vendors may choose different ways to express their equipment functions using the same overall model. Italo - looking for more commonality in how model is used - how to establish best practices? Need to look at how to document best practices.
 
  1. OIF 2018 Demo
Lyndon
  • Vendor survey has indicated at least 8-9 vendors interested in participating
  • Good mix of Domain Controller/NEs and MD Controller/Orchestrators, including one SW vendor
  • All vendors to support Service, Topology, and Notification as well as some support for Path Computation and Virtual Network Service
  • Interest from at least 4 vendors for each Use Case, including all vendors for multilayer setup and rerouting
  • next steps - start on demo tech specs
  • discussion not under NDA at this point, calls are open to OIF and non-OIF members - Thurs 6am Pacific time
  • discussions will become closed after Jan. 12th contracts submission
 

Wireless Transport PoC Results

Giorgio
1) PoC 4.1:
- completed with success
- now in preparation both press release (or media coverage) and white paper
  • hoping before end '17
- Tracy (AT&T) will provide an overview of the PoC during the coming 'ONAP Development Forum' - WT SBI works great with ONAP, able to exchange inventory data with other ONAP component "active and available inventory", able to use SBI with other wireless equipemnt such as remote radio head and IOC centers, dist. antenna system - added interest from equipment vendors to work on an interface to base stations using WT work - hping to bring back more wireless work in ONF - incubate SBI within ONF using Core model, verify/validate with ONAP - SDN-R project in ODL - Tracy will present on this at ONAP
  • note Nigel will be at ONAP as well on Thursday to discuss IM - arriving Wed. PM
  • Lyndon - to provide intro remarks to ONAP developers forum - 10 min or so - Tues. 11:45 am Pacific time - Tracy to provide details
2) Model:
- ongoing review of proposals for model improvement. Target is to have a new version of the model
including ETH PHY within 1Q2018
3) PoC 5:
- starting to collect interest
- Target mid 2018
- not yet defined the sponsor operator.
Important note for OTCC TST: as WT we are using the model (and the Core Model in particular) in the equipment as
southbound interface from controller. This implies that for real implementation in field, a migration from a version to
the next one of the model should be backward compatible to avoid to scratch database and cause long outservices.
In our view we should define at OTCC level a policy for models and in particular Core Model implementation insuring backward compatibilities.
Example:
- not change the type of an object/attribute from version x to version y
- not change an enumerated list (only add)
Nigel: bw compat always a goal, but should take into account lifecycle stereotype in the model, e.g., "experimental" - these should be speculative parts of the model that are more likely to change; "preliminary" much less likely to change but still possible; "mature" or unmarked should be very unlikely to change - would use multiple releases to change with warning to implementors
Karthik - things marked "experimental" should be looked at carefully to provide feedback to the modeling team
Bernd - "experimental" - not mapped to YANG? check on if there is a rule - TAPI uses "preliminary" if YANG is needed? Karthik - change is related to prototyping which will bring greater maturity.
 

ONOS/CORD Activities

 
  • Open Disaggregated Transport Networks project – appears to be committed to addition of TAPI as Northbound Interface to ONOS as part of the project, with NTT leading work
  • XOS – initial work on mapping of information model - mapping to protobufs - Kam/Nigel: work initiated, Chris Hartley has provided initial input and work is being discussed in the IM team. Comments being solicited from CORD folks. Bernd asks for any slides describing this - check for what is available - add as topic for Thursday call
  • Progress in the ITU-T meeting last week, including OAM model - further discusson in OTIM call
 future meetings and other businessLyndon

Next OTCC TST call would be 12/25 - postpone until 1/8

Reminder of doodle poll for March f2f meeting - please sign up if planning to go (looks like WT will participate - question on whether they are planning on meeting as a separate group or just with IM discussion)

  • Old Action Items

  • @Giorgio to provide AAI information and we should aim for follow up discussion on inventory/physical model
  • Scott to provide links to MEF TAPI PoCs
  • Lyndon to check to make sure OTCC TST call is listed on wiki calendar
  • Giorgio to check with WT team on their participation, esp. March or June meeting
  • All: find person to be contact with XOS
  • All: find person to be contact with TIP
  • Action Items

  •  All: think about best practices documentation
  • Kam to check on WT participation in March f2f meeting
  • Lyndon to recheck calendar link on home page, fix if needed
  • Kam to create location for recordings on Tencent