Re: [Isis-wg] Proposed isis-wg documents - draft-ginsberg-isis-route-preference-00

[Uma Chunduri <[e-mail]>]

Les,

Thank you for the response and the complement!

Section 2 of you document re-iterating the corner case where a down bit can be set in L2 LSP.
Ok, some implementation as seen by Stephane did indeed set the "Down" bit in L2 LSP (legit).

But I didn't get answer to my question for TLV 135 as described in Appendix A (Problem) - 
	
	Please clarify, which document said exactly to prefer a route "which does 
	NOT have up/down bit set".


>We also have the existence proof from Stephane's interoperability testing that DOWN bit  is set in L2 LSPs (at least by some implementations) and that interpretation of the DOWN bit in L2 LSPs is not consistent among all implementations. 

I see no specification (5302/5305) said to prefer/not-prefer the route with "down" bit set in L2 LSP. As I said, to me this is implementation issue on R2 not protocol issue for TLV 135.

This has to be clarified before discussing any solution (or position as clarification as you alluded to).

--
Uma C.


-----Original Message-----
From: Les Ginsberg (ginsberg) [mailto:[e-mail]] 
Sent: Wednesday, August 20, 2014 8:40 PM
To: Uma Chunduri; Christian Hopps; [e-mail] list
Cc: Hannes Gredler
Subject: RE: [Isis-wg] Proposed isis-wg documents - draft-ginsberg-isis-route-preference-00

Uma -

The answers to all your questions are already in the draft.

Please reread Section 2 more carefully. It has direct quotes from RFC 5302 and 5305 which state:

" ignore the up/down bit in   L2 LSPs"

And 

" If a prefix is advertised from one area to another at the same
   level, then the up/down bit SHALL be set to 1."

This clearly introduces the possibility that DOWN bit can be set in L2 LSPs AND that it should NOT be used ("ignore" == "don't use"). We also have the existence proof from Stephane's interoperability testing that DOWN bit  is set in L2 LSPs (at least by some implementations) and that interpretation of the DOWN bit in L2 LSPs is not consistent among all implementations. This indicates that clarification is needed - which is the primary purpose of this draft.

The fact that you - a person with significant experience with the protocol - is uncertain also reinforces the need for this draft. (I mean that as a compliment.)

   Les

> -----Original Message-----
> From: Isis-wg [mailto:[e-mail]] On Behalf Of Uma 
> Chunduri
> Sent: Wednesday, August 20, 2014 8:04 PM
> To: Christian Hopps; [e-mail] list
> Cc: Hannes Gredler
> Subject: Re: [Isis-wg] Proposed isis-wg documents - 
> draft-ginsberg-isis-
> route-preference-00
> 
> We know  multiple  places  in 5302 and 5305 it's clearly said "Down" 
> bit MUST NOT be set in L2.
> But I see the corner case in Section 4.1  where "Down" bit was allowed 
> in the same level.
> 
> Now if that is the premise of this work - let's focus bit more on the 
> problem see as described in Appendix A;
> 
> " o  R2 has two possible paths to reach 10/8, Level 2 route with metric
>       2002, up/down bit is 0 (from R0) and Level 2 route with metric
>       101, up/down bit is 1 (from R3).  R2 selects R1 as nexthop to 10/8
>       because it prefers the route which does NOT have up/down bit set."
> 
> 
> Why?
> Please clarify, which document said exactly prefer a route "which does 
> NOT have up/down bit set".
> 
> I could not see it neither in 5302 nor in 5305. To me this is R2 
> implementation issue (NOT a protocol issue) in dealing with 135s.
> 
> 
> --
> Uma C.
> PS: Thanks both Hannes and Chris to check back if my question (asked 
> partially at Toronto) has been addressed.
> 
> -----Original Message-----
> From: Isis-wg [mailto:[e-mail]] On Behalf Of Christian 
> Hopps
> Sent: Tuesday, August 19, 2014 4:48 AM
> To: [e-mail] list
> Cc: Hannes Gredler; Christian Hopps
> Subject: [Isis-wg] Proposed isis-wg documents.
> 
> Hi Folks,
> 
> During the last meeting in Toronto there was good support for making 
> the below drafts WG documents. Unless objections are posted to the 
> list by Friday we will move forward with doing this.
> 
> Proposed WG Documents:
> 
> 	draft-ginsberg-isis-sbfd-discriminator-00
> 	draft-ginsberg-isis-route-preference-00
> 	draft-litkowski-isis-yang-isis-cfg-01
> 
> Thanks,
> Chris & Hannes.
> _______________________________________________
> Isis-wg mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/isis-wg
> 
> _______________________________________________
> Isis-wg mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/isis-wg