[Teas] Interim follow-up: TE Tunnel discussion

["Davis, Nigel" <[e-mail]>]

Hi all, (I previously sent this to only nmop but should have sent to both teas and nmop)

(please read the text below before reading the attached slides as the text provides some context)

I had a backup slide in the slides that I presented at the interim. I have tweaked that slightly (below) to account for the discussion in the second part of the discussion on TE.

[cid:image001.png@01DAB69C.30AA9510]
The discussion noted that the topology of potential is always represented by RFC8345 where as the “topology” of usage is represented in various different ways (examples noted above). In other bodies the approach to representing the Tunnel etc. is to use a single entity type. There appears to be no reason to use a different base entity to represent the “topological” structure at this level of the hierarchy, in fact in TM Forum, there was a very extensive analysis of the representation of connection oriented cases and connectionless cases (~2007) where we concluded that the same model entity could and should be used in both cases and hence the two different entities we had should be merged into one. This unified the model for all technologies. ITU-T also took the same step around about the same time (G.800 resulted from their analysis). The uniformity is highly beneficial for management/control functions as the navigation can be generalized.

After that analysis we considered further possible convergences and concluded that all topological entities could be converged and by considering the problem as one of layered hypergraphs, then the node, link and connection could all be represented by hyperedges the node was then the touchpoint at the termination point. We did not develop that model further in TM Forum, but in ONF in the core model work, it became apparent that a further generalization was possible. This led to the component-system pattern.

Whilst I am not proposing that we go as far as component-system (yet 😊), I do think we should consider the points Olga made about more consistency should be considered seriously in the Digital map work.

The attached slides offer various different degrees of specialization of base classes (including component-system based solution structures).

If you want more background, I can provide more.

Regards,

Nigel