Ran, > This is a minor, almost trivial, extension to very well-understood Traffic Engineering (TE) mechanisms.
I am not sure if you have read my notes to LSR mailing stating the very same. I do support this ISIS extension. > and disagree that any uninterested operator would need to upgrade anything. Never said that. > Further, the operators I’ve spoken with disagree with the claim that any “massive almost network wide > upgrade” would be required. Well if at least one edge does not sing from the same page you may never prepare links for sleeping. Of course it depends where are unupgraded edges and what traffic passes through them. - - - It seems the contention is that extensions to protocols should be scoped and only treated in an myopic manner. On that one we clearly have a difference of opinions. Cheers, Robert On Wed, Sep 30, 2026 at 10:17 PM RJ A <[email protected]> wrote: > > > > On Sep 28, 2026, at 11:15, Robert Raszuk <[email protected]> wrote: > > > > Hi Ran, > > > > While I do support adoption of this TE extension for ISIS I do not agree > that current discussion can be named as "odd". > > > > See IETF rules and guidelines perhaps work decent to define conditions > when this or that draft is proposing an extension which sits already in the > working group charter. > > > > Here this proposal is defining a completely new technology "power saving > by network auto reconfiguration using TE" and by this alone touches work > residing in number of other working groups in IETF. > > > > On the other hand we can't and should not extend protocol to only fit > one design/architecture as changing such extension while deployed out in > the wild is very costly - especially for architectures which to work well > do requires a massive almost network wide upgrade. > > I believe that you believe that. > > Based on my ~38 years of IETF experience, and having personal experience > with network engineering for a multi-continent broadband ISP, I disagree > with your characterization. > > This is a minor, almost trivial, extension to very well-understood Traffic > Engineering (TE) mechanisms. Further, I disagree that this proposal > narrowly fits “one design/architecture” and disagree that any uninterested > operator would need to upgrade anything. Operators that choose to use the > mechanism are perfectly capable of making their own cost/benefit analyses. > Further, the operators I’ve spoken with disagree with the claim that any > “massive almost network wide upgrade” would be required. > > Yours, > > Ran > > _______________________________________________ > Teas mailing list -- [email protected] > To unsubscribe send an email to [email protected] >
_______________________________________________ Lsr mailing list -- [email protected] To unsubscribe send an email to [email protected]
