Re: shim6 @ NANOG (forwarded note from John Payne)

[Tony Li <tony.li@tony.li>]

Jason,

> As Ted Seely from Sprint said, .Isn.t new technology supposed to be better
> than the technology it replaced?. 


'Better' is certainly open to interpretation and may be evaluated by
differing parties according to different metrics, thus leading to
different results.  One might recall that even since the introduction of
the CD, there is still a steadfast group of audiophiles who feel that
the sound of vinyl is 'better'.


> Is it the consensus of the shim6 working group that the full suite of TE
> capabilities should not be a requirement?  Or is this just the opinion of
> a few vocal people?


I can't speak to the consensus, but so far the requests for TE
capabilities seem to be isomorphic to a request for a return to the v4
routing architecture, complete with PI addressing and global prefix
advertisement.

While I have no objection to providing any TE functionalities as part of
our multihoming solution, I think that we also need to remember that our
solution also needs to satisfy other constraints: namely that we have
tried a PI-based solution before and that we are trying very hard to
avoid that segment of the solution space again due to the cost
implications for the routing subsystem.  We are also operating within
some (possibly undue) constraints imposed by the existing v6
architecture.  Given these constraints, if we add in a requirement to
replicate all of the functionality that we have with v4 TE, we may well
find that the solution space is void.

Thus, I feel that we have to prioritize our constraints.  To me,
creating a scalable routing subsystem is paramount, as without that, we
effectively have no network.  Second, due to the realities of the day,
we need to respect the portions of the v6 architecture that cannot be
changed, at least within this WG.  And lastly, we can provide any
feature functionality that is compatible with the above constraints.

Regards,
Tony