Re: [i2rs] IETF 89 agenda published

["Susan Hares" <[e-mail]>]

Jeff: 

I like, Tom would like to know what the status of the i2rs use-case work?  

The following will be presented at IETF individually:  
http://datatracker.ietf.org/doc/draft-white-i2rs-use-case/ 

These use cases provide the distributed reaction to network attacks, remote
service routing (hub spoke), within data center routing, and temporary
overlays.  
 
http://datatracker.ietf.org/doc/draft-keyupate-i2rs-bgp-usecases/

This has been proposed for the BGP handling case to retrieve information on
the errors, topologies and flowspec, BGP Events, and statistics.  I2RS could
also push information to allow the pushing customer cost communities,
flow-spec routes, new filters for legacy routes, optimize exit control (via
existing policy or PBR).  

I am presenting these following of these as a group for my co-authors: 

http://datatracker.ietf.org/doc/draft-hares-i2rs-use-case-vn-vc/
http://datatracker.ietf.org/doc/draft-huang-i2rs-mpls-te-usecases/
http://datatracker.ietf.org/doc/draft-chen-i2rs-mpls-ldp-usecases/
http://datatracker.ietf.org/doc/draft-ji-i2rs-usecases-ccne-service/
(centralized controller) 
http://datatracker.ietf.org/doc/draft-chen-i2rs-ts-use-case/
(traffic steering) 

This grouping provides a general framework  service layer routing
improvements to hub-and-spoke the i2rs charter requests based on a virtual
networks/virtual connections (draft-hares-i2rs-use-case-vn-vc), utilizing a
centralized controller (draft-i2rs-usecases-ccne-service), mpls-te links
(draft-huang-i2rs-mpls-te-usecases),  mpls ldp
(draft-chen-i2rs-mpls-ldp-usecases), and traffic
Steering (draft-chen-i2rs-ts-use-case).  My co-authors and I seek feedback
on these use cases.

My co-authors and I would love to chat about the mobile backhaul use case: 

http://datatracker.ietf.org/doc/draft-zhang-i2rs-mbb-usecases/

but since it is not specifically on the charter, I suspect you and Ed will
need to rule if can be discussed on the list.  If it is not, please let me
know. 

It is important to get the use cases for these deployments down so we can
adequately discuss the information models. 

Sue Hares

-----Original Message-----
From: i2rs [mailto:[e-mail]] On Behalf Of t.petch
Sent: Friday, February 28, 2014 10:18 AM
To: Jeffrey Haas
Cc: [e-mail]; Edward Crabbe
Subject: Re: [i2rs] IETF 89 agenda published

----- Original Message -----
From: "Jeffrey Haas" <[e-mail]>
To: "t.petch" <[e-mail]>
Cc: "Edward Crabbe" <[e-mail]>; <[e-mail]>
Sent: Thursday, February 27, 2014 4:32 PM
> Tom,
>
> On Thu, Feb 27, 2014 at 02:20:17PM +0000, t.petch wrote:
> > It seems a shame that use cases comes last, since, as Alia said 
> > recently,
> >
> > "If you were to look at our charter, unsurprisingly we have
use-cases to
> > be
> > completed before information models.  I would strongly encourage 
> > discussion of the use-case drafts and serious work on turning them 
> > into
something
> > that
> > the working group could accept."
> >
> > My own take is to wait for use cases to progress before taking a
serious
> > interest in, e.g., architecture and info model.
>
> As I've been coming up to speed on the documents in the work group and
the
> mail archives, the use cases vary from very well discussed to very
light
> discussion.  The ones that have received good discussion also tend to
have
> related info model documents.
>
> Since I've been digesting the use case documents in bulk, it's pretty
clear
> that a number of common requirements can be extracted from them.
Thus,
> where the documents appear to be in terms of maturity is some need of 
> editing and refinement to attempt to make those common requirements 
> explicitly visible across the document set.  Much of that work is for
the
> mailing list.  Some of the use cases are a bit more novel in terms of
the
> likely fallout of their requirements and deserve more group
discussion.
>
> In terms of simply keeping the WG pipeline full, it's definitely time
to
> start spending some cycles on data model language and protocol.  While
it's
> certainly possible that as the use case documents are refined that
they'll
> reveal additional requirements, my hope is that they're not disruptive 
> discoveries.  Clearly it's a finish-to-finish dependency relationship.
>
> Meanwhile, I'm sure the use case authors would appreciate as much help
as we
> can give them to drive the documents to completion. :-)

Jeff

I find the use case documents a very mixed bag.

I note that the rib info model cites the white and hares use case I-Ds,
which suggests to me that they should be adopted by the WG (or else
dropped:-(

I note that the mbb usecases, which I find the most persuasive, regards the
info model as inadequate, which, again, says to me this should either be a
WG document, or else declared out of scope for the info model at this time.

Discussions last year seemed to stall at times over what the use cases
should be and I would find it hard to know whether or not the info model was
adequate without knowing what the use cases were.  So I think that the WG
needs a view on which of the many use case I-Ds is relevant to the current
work.

I find the protocol issues less pressing, perhaps because the IETF has been
round that block so many times before.  The genesis of e.g. PCEP seemed a
relatively straghtforward birth, compared to, say, the data modelling work
of e.g. netmod or, delving into the past, SMI.

Tom Petch

> -- Jeff (not currenty speaking as a chair, just someone who's done
some
> amount of project management)
>

_______________________________________________
i2rs mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/i2rs