Uma -
I share the concerns expressed by Ketan and Peter.
Although I will certainly consider the additional response you seem to have
hinted at in your reply to Peter, it seems to me that Section 6 of your draft
acknowledges that there is a scaling problem - and then references what seems
to be a non-existent draft (I could not find "draft-ce-ppr-graph-00" ???) as a
proposed solution.
In any further response you make it would be good if you did indicate whether
you agree PPR has a scaling issue - and given the contents of Section 6 of your
draft - also discuss why we should consider PPR + some additional scaling
enhancements when we already have at least one solution which does scale
linearly.
Thanx.
Les
> -----Original Message-----
> From: Lsr <[email protected]> On Behalf Of Uma Chunduri
> Sent: Tuesday, July 17, 2018 6:43 AM
> To: Peter Psenak (ppsenak) <[email protected]>; Ketan Talaulikar (ketant)
> <[email protected]>; [email protected]
> Cc: [email protected]
> Subject: Re: [Lsr] Concerns with
> draft-chunduri-lsr-isis-preferred-path-routing
>
> Hi Peter,
>
>
> -----Original Message-----
> From: Peter Psenak [mailto:[email protected]]
> Sent: Tuesday, July 17, 2018 9:34 AM
> To: Uma Chunduri <[email protected]>; Ketan Talaulikar (ketant)
> <[email protected]>; [email protected]
> Cc: [email protected]
> Subject: Re: [Lsr] Concerns with
> draft-chunduri-lsr-isis-preferred-path-routing
>
>
> > flex-algo is NOT a per path state. It's a regular prefix-SID, which is
> bound to a prefix only.
>
> Not quite. I will come back to this bit later.
>
> >The problem with your proposal is that you flood all paths to all
> routers
> in the area and ask every router to evaluate all of them and see which of them
> may apply. This scale poorly.
>
> Please see Section 6, in 01 version.
>
> --
> Uma C.
>
> _______________________________________________
> Lsr mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/lsr
_______________________________________________
spring mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/spring
