Time | Item | Who | Notes |
---|
10 mins | Administrative |
| |
60 mins | Intent / Co-routing | Andrea Mazzini | - Discuss Identification of "intent" modeling items #432
- Andrea shows the TAPI Connectivity Diagram, discussion on the behavior of _corouteInclusion of ConnectivityConstraint composed class of ConnectivityService.
 - Discussed some possible behaviors, given ETH_1 the ConnectivityService under provisioning, and ETH_0 the ConnectivityService indicated by _corouteInclusion attribute.
- Snapshot, i.e. inclusion (and exclusion) is performed only at first routing of ETH_1, no further dynamic alignment to any route change of ETH_0
- Fate sharing, i.e. ETH_1 will follow any ETH_0 rerouting
- Strict association, i.e. ETH_1 and ETH_0 share lifecycle, when ETH_0 is deleted, also ETH_1 is deleted and vice-versa
- Loose association, e.g. in LAG/LCAS scheme
- Kam suggests that the issue shall be explored also in Core IM
- other possible scenarios, the ETH_0 route is either shorter (possible sequential concatenation/stitching) or longer than ETH_1
- Malcolm: evaluate the relationship with the "bundled Connectivity Services" as proposed for photonic model
- Nigel: concerning TAPI, there are two simple solutions:
- Specify that ETH_0 cannot be deleted until ETH_1 is deleted, so preventing empty references
- Specify that a null pointer may appear in case ETH_0 is removed. Original intent (ETH_0 id) should be found in the log.
- Nigel: all these different behaviors shall be specified through policy, Andrea to check MEF documentation on the subject.
|
30 mins | Partitioning & Abstraction | | - Continue discussion started during 2019-06-03 TAPI Workshop Notes, on updated presentation
- According to suggestions from previous call, added a couple of "bottom-up" scenarios, oriented resp. to equipment and routing views.


- Agreed that "physical" resources can be partitioned, aka slicing.
- Clarified that the purpose of these slides is the exploration of possible enhancements / generalization of current "semi-opaque" topology model.
Use cases shall drive future decisions.
|
| Photonic model, “bundled” Connectivity Service, | Not discussed as key people missing |
|