No Subject

[Jim Boyle <jimpb@nc.rr.com>]

Angela (and John) - this is indeed an interesting paper to review,
thank you for posting it (indirectly) to the list.  Here are some
comments.

I don't think this impacts the current restoration and hierarchy
requirements.  No one is denying that in the future, great things will
be possible by seamlessly merging optical and router domains (ok, so a
few are, but that's besides the point :) The design team decided not
to highlight the inevitible work in this area for several reasons,
such as:

     -) no one has quite come to the point where they are having
        problems interfacing their GMPLS capable all optical core with
        their massive OC48/192 networks where the obvious solution is
        to somehow interact at the network layer.

     -) in fact, one might speculate that *at this time* no one has a
        IP capaple WDM core.

     -) there *does* appear to be a pressing need in the area of
        supporting VPNs, especially those with associated SLAs, in a
        scalable manner, which might require some hierarchy.

     -) on restoration, there is a pressing need in data networks as
        well as TDM networks (sufficiently decoupled from their data
        bretheren) to have a small set of interoperable approaches at
        service restoration (a path based one, and something
        "quicker").

So for now, at least, the thought is to focus on pressing needs and
for now, at least, try not to muddle the solutions with rather forward
looking architectural problems.  Once the pressing needs are
addressed, then it might be again time to revisit needs for
integration of TDM and IP networks (who sponsors those requirements is
not defined right now, though).

I'm up for hearing feedback on this, but to date, this has been the
stated position and there hasn't been much objection, at least from
those whose responsibilities include running networks.

tick tick tick, though, the ADs say the timers on this are passing...

regards,

Jim

p.s. I have some technical feedback to the document - should it be sent to
you and John - or to the IPO list??

On Wed, 5 Sep 2001, Angela Chiu wrote:

> Hello all,
>
> A couple of weeks ago, we sent the authors of the restore-hierarchy
> draft the following email to consider the proposed Joint IP/Optical
> Restoration Scheme for Router Failures as part of the survivability
> requirements that are specific to the IP over Optical architecture with
> the paper attached (which we could not send to the general WG mailing
> list). Now that we have managed to put it on a public accessible website
> http://www.celion.com/tech.html, we would like to share it with the
> members of the TE community to receive your comments and suggestions.
>
> Note that the simple network topology chosen in the paper is for
> simplicity in illustrating the restoration mechanisms. The schemes can
> provide the same level of effectiveness for more general topologies,
> such as a mesh network with more backbone links between offices as
> described in the 2nd to the last paragraph in the paper, and cases where
> there are more than one backbone links between two offices connecting
> either the same 2 backbone routers or different ones.
>
> Best regards,
>
> Angela Chiu
> John Strand
>
> -----Original Message-----
> From: Angela Chiu
> Sent: Friday, August 24, 2001 6:11 PM
> To: 'wlai@att.com'; 'dave.mcdysan@wcom.com'; 'jimpb@nc.rr.com';
> 'malin@sunet.se'; 'rcoltun@redback.com'; 'griffin@research.att.com';
> 'ejk@tech.org'; 'treddington@bell-labs.com'
> Cc: 'James Luciani'; Awduche, Daniel; John Strand (E-mail 2); Angela
> Chiu
> Subject: RE: [IP-Optical] draft-team-tewg-restore-hierarchy-00.txt:
> Request for input on restoration requirements
>
>
> Dear authors of the restore-hierarchy draft:
>
> Since the IPO WG is asked to provide inputs on requirements specific to
> restoration in the IP over Optical architecture, we thought the attached
> work should be relevant to this topic. It came directly from Worldnet IP
> network on the real problem they have been facing -- the mismatch
> between the restoration requirement from IP networks and the restoration
> capability at the transport layer. On one side, fast
> protection/restoration can be provided by the optical/SONET layer to
> have protected links in an IP network. On the other side, IP networks
> need to put in additional backbone capacity to account for every single
> router failure to avoid congestion during a router failure. In AT&T
> case, it sometimes can be a substantial percentage of the total backbone
> capacity. Since this extra capacity can also cover for any single link
> failure, there is less incentive to buy protected links especially when
> they cost twice as much as unprotected ones. Our proposal below provides
> a way to solve this problem.
>
> We thought it might be good to send you this to be considered for the
> restoration hierarchy draft to address some needs specific to the IP
> over Optical architecture. (Note that the paper described the motivation
> and high-level functional requirements on the mechanisms. Detailed
> protocols still need to be developed by other WG such as CCAMP if this
> does become one of the requirements.)
>
> If you have any question regarding the work, feel free to call us. Your
> comments/suggestions will be greatly appreciated.
>
> I will also send the paper to the TEWG mailing list once I find a public
> website to put it since IETF does not allow people to send the actual
> document to the mailing list.
>
> Best regards,
> Angela Chiu
> Celion Networks
> 1 Shiela Drive, Suite 2
> Tinton Falls, NJ 07724
> Tel. (732) 747-9987
> Fax (732) 747-9986
> email: angela.chiu@celion.com
>
> John Strand
> 510 642-9719 (W)
>
>
> -----Original Message-----
> From: James Luciani [mailto:jluciani@CrescentNetworks.com]
> Sent: Friday, August 17, 2001 8:55 AM
> To: ip-optical@lists.bell-labs.com
> Cc: jboyle@worldview.com; Awduche, Daniel
> Subject: [IP-Optical] draft-team-tewg-restore-hierarchy-00.txt: Request
> for input on restoration requirements
>
>
> Folks,
>    The TE working group, as per the request of our ADs, is requesting
> input
> from IPO on
> http://www.ietf.org/internet-drafts/draft-team-tewg-restore-hierarchy-00
> .txt
> from an IP over Optical perspective.   Please give the draft a read and
> give
> your input ASAP as they would like to move the document forward next
> month
> barring any substantive input.
> Thanks for your cooperation and input in advance.
> --Jim
> P.S., draft can be found at
> > Hello Jim,
> >
> > As we discussed in London, and as many are aware, the TE-WG is working
> on
> > a requirements document in the areas of restoration and hierarchy.
> The
> > goal is to be identify the focussed set of requirements that service
> > providers are challenged with at this time.
> >
> > The draft in question has been adopted as a TEWG item, however its
> > content is being actively discussed on the TEWG list, and it will be
> > revised.
> >
> > What the design team determined was that at this time the following
> are
> > most pressing:
> >
> > - a small, consice set of restoration approaches for initial
> >   standardization and interoperability.
> > - one path based approach
> > - one "local repair" approach
> >
> > - focus on hierarchy in the context of edge-to-edge signalling
> >           requirements for networks with multiple IGP areas (e.g. OSPF
> >           areas, ISIS levels) or flat IGPs of large size (and thus,
> >           raising concern of scalability limits and ways to extend
> >           these.  The percieved driver for this work is VPNs with SLAs
> >           in data networks.
> >
> > Interestingly enough, inter-layer hierarchy, that is communication and
> > coordination between WDM/SONET layers and Router layers, was not
> > determined to be as interesting at this time.  This is from the
> > perspective of operators who wish to use and deploy this technology,
> > as opposed to protocol developers who see the natural progression and
> > want to flesh out the details.  Main issues discussed included
> > interoperability and administrative issues with dynamic physical layer
> > networks *at this time*.
> >
> > I solicit input on the requirements from members of the IPO WG either
> > directly as members of TEWG, or in some consolidated fashion
> > (e.g. bring results of discussion on IPO to TE).
> >
> > We have commited to attempt to get these requirements over to CCAMP by
> > September 30, and would like to have all new comments to the TE list
> > by no later than the September 7th, if possible.  Feel free to use
> forward
> > this message onto IPO, or to use other means of notification as you
> and
> > Dan see fit.
> >
> > regards,
> >
> > Jim
> >
> >
> >
>
>
>
> _______________________________________________
> IP-Optical mailing list
> IP-Optical@lists.bell-labs.com
> http://lists.bell-labs.com/mailman/listinfo/ip-optical
>