Re: Status update about GIST

john, 

> -----Original Message-----
> From: [email protected] [mailto:[email protected]] On 
> Behalf Of [email protected]
> Sent: 28 May 2008 03:19
> To: [email protected]; [email protected]
> Subject: Re: [NSIS] Status update about GIST
> 
> Martin,
> 
> Does this mean that the use RAO is basically thought of as a bad thing
> by the IESG?

my interpretation of the discussion is that: 

IF your protocol
* uses RAO, and
* has any likelihood of significant deployment on the Internet
  (where "significant" is defined as "more than RSVP for QoS")
THEN 
* it will be experimental

... at least for IPv4. the IPv6 case is (at least to me) much
less clear cut: theoretically/architecturally they should be
the same, but the practical/engineering issues are quite different.

one of the routing ADs is going to write a draft explaining this.

robert h.

> 
> John
> 
> -----Original Message-----
> From: [email protected] [mailto:[email protected]] On 
> Behalf Of
> ext Martin Stiemerling
> Sent: Tuesday, May 27, 2008 6:36 PM
> To: nsis
> Subject: [NSIS] Status update about GIST
> 
> Hi all,
> 
> It has been a bit silent the last weeks about what is 
> actually happening
> with GIST. You might still remember that GIST go sent to the IESG and
> underwent the IESG review and discussion. 
> 
> The GIST spec got quite an amount of ABSTAINS during the IESG 
> review and
> therefore is blocked. Most of the IESG members have had concerns about
> GIST and this led to the ABSTAINS. 
> 
> However, to get things progressed, Magnus initiated a joint telechat
> between the IESG, Robert, Henning and me to discuss the ABSTAINS. The
> call took place on Thursday, May 15, 5:30 pm CEST. Here is 
> brief summary
> (that reassembles my memory): 
> 
> The discussion with the IESG centered around the Router Alert 
> Options in
> IPv4 but also touching IPv6. The general concern about thiss is that
> NSIS could lead to a flood of IP packets including IP RAO options and
> that this could effect existing RSVP-TE deployments, plus all current
> routers (as IP options go to the slow path), and finally that packets
> with RAO options are anyhow filtered at the edges of the 
> Internet. This
> cause most of the ABSTAINS and the call for GIST being pushed to the
> experimental track instead the standards track. 
> 
> The issues associated with RAO were discussed for quite some while (I
> remember 1 hours or so) and we couldn't find any agreement 
> for the IESG
> on staying with RAO as part of the GIST spec AND going for standards
> track. The envisioned solutions to this is, revert to packet filtering
> mechanism (e.g. 5-tuple) to intercept GIST packets, instead of RAO, in
> the GIST specification. This would be allowed to go GIST as standard
> track. The RAO part will be cut out of the GIST spec and pushed to a
> separate document. This new document will go on experimental track. 
> 
> This splitting will also effect the NSLPs, as they are also 
> definin RAO
> usage. 
> 
>   Martin
> 
> 
> 
> [email protected]
> 
> NEC Laboratories Europe - Network Research Division NEC 
> Europe Limited |
> Registered Office: NEC House, 1 Victoria Road, London W3 6BL |
> Registered in England 2832014
> _______________________________________________
> nsis mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/nsis
> _______________________________________________
> nsis mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/nsis
>