Sixty-ninth IETF Chicago July 2007

Tuesday, July 24, 2007
0900-1130 Morning Session I
RTG     ccamp     Common Control and Measurement Plane WG

CCAMP Working Group Minutes, Session 1 of 2
============================================
Chairs: Adrian Farrel 
        Deborah Brungard 

Note taker: Michael Dueser

================================================
0. Administrivia (chairs, 5, 5/150)
   Slides

Adrian opened, gave Note Well, noted that CCAMP will meet twice.
Asked for comments on the agenda. No comments.
================================================

================================================
1. WG status, RFCs, drafts, milestones, charter (chairs, 15, 20/150)
   Slides

4 new RFCs, 3 in Editor's Queue, 5 in AD/IESG review.
On the Charter, the RSVP-TE extensions MIB is overdue.

Tomohiro Otani - Volunteered to start work.
Adrian - Thanks! please coordinate with me for ideas about what needs to
be done.

Zafar - The authors of mpls-graceful-shutdown believe that the
I-D is ready for WG Last Call.
Adrian: OK.

Adrian reviewed proposal for adding GELS work to Charter. Will continue
the discussion tomorrow during the 2nd session.
================================================

================================================
2. ITU-T and OIF progress report (Lyndon, 5, 25/150)
   Slides

Lyndon presented slides. OIF met in April 2007. Planning a demo for
2007 (ECOC). Doing VCAT/LCAS and not using CCAMP's approach. Making some
use of LSP Hierarchy bis draft. Liaisoned to CCAMP (this meeting) on
OIF's interlayer model as input to VCAT work. Some aspects of the model
prototyped for demo. On ITU, 4 liaisons to CCAMP.

Adrian - There was an ITU/IETF leadership meeting on Saturday where the
Q14-CCAMP work was discussed. As Q14 raised the concern that they are
uncomfortable with some of the text in the ASON RFCs, we need to liaison
them to Q14 to understand what are their specific concerns.
Emphasized that more detailed input is required to understand their concerns.
================================================

================================================
3. ASON Routing Solution (Dimitri, 15, 40/150)
   - Status and plans
   Slides

Background reading
   - draft-ietf-ccamp-gmpls-ason-routing-ospf-03.txt

Dimitri said this is ready for working group last call. Asked if there
were any implementations? (no one answered)

Adrian - We have the responsibility to do this work, and this is the
only I-D in this space, so we will ask the mailing list.

Dimitri - We will have a face-to-face meeting on Wednesday at lunchtime
to try to close on some of the open issues on ASON routing especially
the issues raised by Q14.

Adrian - The scope of the current draft is to satisfy the RFCs. Updates
to the RFCs might be made after they have been liaised, and then we can
revisit the solution

================================================

================================================
4. VCAT/LCAS (Greg, 15, 55/150)
   - Progress and plans
   - Issues to be resolved
   Slides

Background reading
   - draft-ietf-ccamp-gmpls-vcat-lcas-02.txt

Lyndon - ID is helpful in describing terminology, right approach in
going forward to incorporate OIF and ITU concerns.

Igor - What is the VCG layer?
Jonathan - It's an intermediate layer used for further separation and
clarification. VCAT is the client layer to the STS-layer, then used to
run Ethernet over STS.

Adrian - Danger of divergence; need to obtain one implementation spec
for the entire industry (IETF, OIF + ITU)
Lyndon - The prototype developed by OIF is not a standard, wish to work
on a common solution.

Adrian - We should make sure to liaise ITU the latest VCAT I-D.
================================================

================================================
5. Multi-Layer Network (Dimitri, 10, 65/150)
   - Progress and plans
   - Issues to be resolved
   Slides

Background reading
   - draft-ietf-ccamp-gmpls-mln-reqs-03.txt
   - draft-ietf-ccamp-gmpls-mln-eval-03.txt
   - draft-papadimitriou-ccamp-gmpls-mrn-extensions-04.txt

Dimitri asked that the requirements and evaluation draft be last called,
and the solution draft be considered for working group status.

Lyndon - Not all issues raised in the liaison have been considered
(especially the multi-layer client-server relationships, scalability
issues).
Jonathan - Different units within the same company may operate on
different network layers.
Dimitri - The ID was never meant to provide a generalized model covering
all situations.

Jonathan - Need to separate between interworking of protocols and
interworking of architectures.

Adrian - Suggests to liaise back to the ITU, and to consider comments as
part of the IETF WG last call process.
Jonathan - Not enough time for ITU to respond.
Adrian - ITU could respond to liaison through correspondence as did on
previous liaisons. We'll take the extensions draft to the list for
consideration as working group document.
================================================

================================================
5a. IGP Extensions for Inter-AS TE Links (Mach, 10, 75/150)
   Slides

Background reading
   - draft-ietf-ccamp-ospf-interas-te-extension-00.txt
   - draft-chen-ccamp-isis-interas-te-extension-00.txt

Adrian - We need to continue to consult carefully with the IGP working
groups. We should make the IS-IS I-D a working group draft.

Dave Ward - Will you coordinate with the OSPF and IS-IS mailing list?
Adrian - Yes.
================================================

================================================
6. Conversion between Permanent Connections and Switched Connections (Dan, 10, 85/150)
   - Progress and plans
   - Solutions?
   Slides

Background reading
   - draft-ietf-ccamp-pc-and-sc-reqs-00.txt

Dan presented the slides, asked if should merge the solution draft with
the requirements.

Adrian - Should combine or separate? (most said to keep the documents
separate)

Dave M. - This is important work. Want to help with requirements. Look
out for corresponding MIB updates!
================================================

================================================
7. ARP over GMPLS (Zafar, 15, 100/150)
   Slides

Background reading
   - draft-ali-arp-over-gmpls-controlled-ethernet-psc-i-03.txt

Dimitri - Need to be clear which issues are to be discussed in CCAMP,
and which are more general issues of IP address resolution over Ethernet links.
The proposed solution is trying to solve this issue via the control plane used
for establishing the Ethernet link but this is orthogonal to the present problem.
In this case you would like to ensure that address resolution does not make use of
the tunnel end-point IP address (associated to the control plane) used for
establishing the LSP between the routers.

Adrian - Need to understand the problem properly - don't see major
problems but it's fine to document.

Zafar - Want to document a number of issues which emerged during
inter-op tests.

Adrian - Need to clarify a number of issues, discuss on the list.
================================================

================================================
8. OAM Requirements for GMPLS Networks (Tomohiro, 15, 115/150)
   Slides

Background reading
   - draft-nadeau-ccamp-gmpls-oam-requirements-02.txt

Adrian - This work is on the charter, ready to be adopted if the WG
agrees. Concerned that 90% of draft is a list of requirements that
seem to have been met already. Need to identify additional operational
requirements. Requirements should be pre-requisite for solutions work.

Dimitri - More input from service providers would be helpful.
================================================

================================================
9. MPLS/GMPLS Security Framework (Luyuan, 15, 130/150)
   Slides

Background reading
   - draft-fang-mpls-gmpls-security-framework-01.txt

Adrian - This work is being done in the MPLS WG, so CCAMP
must monitor and contribute there.
Are there any burning issues to date? How does the I-D affect the
current work of CCAMP?
Luyuan - No, the sky is not falling on our heads.

Dimitri - You need to consider what is the security impact of using IP
addressing space for non-IP terminating interfaces (data plane)? I will
comment on the list.

Zafar - Consider splitting document, general and details by technology?
================================================

================================================
10. Applicability of GMPLS and PCE to wavelength switching (Greg, 10, 140/150)
   Slides

Background reading
   - draft-bernstein-ccamp-wavelength-switched-00.txt

Zafar - Is it signaling or routing that have problems?
Greg - Signaling is in good shape, the problem is with routing.
Zafar - What is needed?
Greg - Currently we don't know which wavelength is available on what
links.
Zafar - Different approaches could be used, need to consider scaling.
Dimitri – Spectral routing constraints are to be considered in combination
with spatial routing constraints, dynamic routing information exchange in
such environment result in scalability and stability issue on the IGP.
================================================

================================================
11. Lambda Labels (Tomohiro, 10, 150/150)
   Slides

Background reading
   - draft-otani-ccamp-gmpls-lambda-labels-00.txt

Adrian - There is a sudden flurry of work appearing in this area - so
there is interest.
Looks like we can do two pieces of work at once:
   1. A framework to scope the problem (Greg's draft)
   2. A specific solution (in a small draft) for lambda labels
   Other work may follow

Dimitri - is this targeted at the IGP?
Greg - Not targeted anywhere. This is to understand what information
is needed. When we know that we can figure out how to distribute it.

Don O'Connor - Support this work, but not ready for WG status yet.
================================================



************************************************

Wednesday, July 25, 2007
1300-1500 Afternoon Session I
RTG     ccamp     Common Control and Measurement Plane WG

CCAMP Working Group Minutes Session II
=====================================
Chairs: Adrian Farrel 
        Deborah Brungard 

Note takers: Michael Dueser and Martin Vigoureux
================================================
12. Administrivia (chairs, 5, 5/120)
   Slides

Wataru's draft was added as related to other drafts.
================================================

================================================
12a. Routing Extensions for Switching Constraints (Wataru, 5, 10/150)
   Slides

Background reading
   - draft-imajuku-rtg-sw-constraint-02.txt

Adrian - Is the advertisement static or dynamic?
Waturu - Dynamic.

Adrian - Fits with Greg's work for consideration.
================================================

================================================
13. MEF UNI (Lou, 20, 30/120)
   Slides

Background reading
   - draft-berger-ccamp-gmpls-mef-uni-00.txt

Lou commented planning to split the document (1) describe UNI specific
(small draft) (2) describe MEF services. After respin, would like to
ask for working group status.

Adrian - I will look at how we can communicate to the MEF direct
Adrian -  I have some questions that I will take to the list

Dimitri - On EPL, what switching type (number)? If below L2SC, even TDM,
need to describe the relation to other switching types.
Adrian - Send comments on the list. We'll need to liaise to other SDOs,
will look at how to communicate with MEF.
Lou - We are informally talking to other organizations, especially the
OIF. Labels seem to be in line, let the chairs decide when to liaise.

Dimitri - Consider whether you have/need a bidirectional LSP or a
"bi-uni-directional" LSP (i.e. a pair of uni-directional LSPs)
Is the label really symmetric? Should we say downstream inferred?
Lou - No problem with changing the name.

Dimitri - Should investigate forward compatibility, should support
asymmetrical UNI e.g. if only one side supports UNI?
Lou - OK

Julien - I support the work, it fills a gap. In case of a large amount
of VIDs, you propose to use multiple LSPs but shouldn't this be decided
by the transport?
Lou - The service interface is represented by a call. From a
signaling standpoint we may signal it by multiple lsps. So we are in
sync that it is up to the provider to decide.
Adrian - Please be careful to distinguish control plane vs. data plane
LSPs. The discussion here is about whether to use one or more control
plane LSPs.
Julien - Please make clear that "number of LSPs" refers to signaling
    state and not to data plane state

Julien - How to co-ordinate VLAN IDs at both ends of the service?
Lou - There is an option to use the same VLAN IDs. It's more a data
plane issue which is outside the scope of this group

Dimitri - Suggest to clarify this. I will help with text on "LSP" so
there is no confusion, especially if liaise it.
================================================

================================================
14. Bidirectional Lightpath signaling (Sugang, 10, 40/120)
   Slides

Background reading
   - draft-xu-rsvpte-bidir-wave-00.txt

Dimitri – You do not need to carry both upstream and downstream label set,
if you develop a mechanism which operates within the existing label set object
nd allow for selection of the upstream within that set (by using e.g. LSP
attribute flag) you could greatly simplify this I-D.
Adrian - Need to taken to the list for further discussion
Adrian - Is it possible to obtain a generic solution for symmetric
labels (Ethernet and wavelengths)? Should try to find a generic
solution.
Lou - Agree with Dimitri. Will work for a generalized solution.
================================================

================================================
15. MEF Ethernet Traffic Parameters (Dimitri, 10, 50/120)
   Slides

Background reading
   - draft-ietf-ccamp-ethernet-traffic-parameters-02.txt

Dimitri commented that from discussion with MEF participants, TLV field processing
has been detailed for mCoS EVC in the new release. He will respin the draft soon to
reflect MEF UNI when stable (sync with Lou).
================================================

================================================
16. GELS-related (Don, 15, 65/120)
   Slides

Background reading
   - draft-fedyk-gmpls-ethernet-pbb-te-01.txt

Ali - Should refer to the IEEE document instead of having an
informational I-D covering the definition of PBB-TE.
Ali - What is the purpose and objective of another liaison on QinQ? Do
we want to do anything beyond PBB-TE?
Adrian - A liaison is required only if IEEE has concerns about work in
CCAMP. We need at least to ask IEEE to see if they have any concerns.
George Swallow - Supports a liaison to the IEEE considering the issues
with the ITU/IETF T-MPLS interactions.
Loa - The IESG instructs us to liaise if control plane work is carried
out for IEEE data plane technologies

Florin - It will be difficult to keep an informational RFC in sync with
the changes expected in PBB-TE formats, standards.

Dimitri - What does Ethernet types mean? Concern that generalized control plane
is achieved (with GMPLS), but this behavior will change with technologies that
are not following generalized operations performed by LSR e.g. PBB-TE switch
does not define an LSR.
Don - .1Q and PBB TE.
Dimitri - Are we looking at a label distribution protocol or a resource
reservation protocol only? We need to answer the question before progressing.

Vach Kompella - We ought to do a liaison to IEEE. How should CCAMP react
in case IEEE creates its own control plane?
Ali - IEEE is doing control plane.
Don - Yes, but no overlap (not TED).
Adrian - Need to look at that issue very carefully.
Zafar - There is functional overlap.
Kireeti - In IEEE, Don, you have a document TE for Ethernet networks
using ISIS. Do we have to interwork with this?
Don - This is for shortest path trees, no need to interwork.

Dimitri - Could we have a framework draft before moving on? Will there be
operational issues within
its scope?
Adrian - We could, authors should propose a split for the current work.
George Swallow - Should pay attention to Qat, IEEE is addressing
reservations.
================================================

================================================
17. Asymmetric Bandwidth Bidirectional LSPs (Lou, 15, 80/120)
   Slides

Background reading
   - draft-berger-ccamp-asymm-bw-bidir-lsps-00.txt

Lou suggests that this should be an experimental I-D so that both
technical options can be kept open, invite feedback.
================================================

================================================
18. LSP Dynamical Provisioning Performance Metrics in GMPLS Networks (Dan, 10, 90/120)
   Slides

Background reading
   - draft-xie-ccamp-lsp-dppm-01.txt

Adrian - Question if this work is within the CCAMP charter?
Is the work necessary? (10 hands - mostly SPs)

Jean-Louis - Needed, but similar work in the Benchmarking WG.
Adrian - This will also be considered in the APM BoF after this session.
================================================

================================================
19. Performance Analysis of Inter-domain Path Computation Methodologies (Jau, 10, 100/120)
   Slides

Background reading
   - draft-dasgupta-ccamp-path-comp-analysis-00.txt

Dimitri - Need to distinguish the different crankback methodologies when
carrying out comparisons and need more precisely to define the notion of
failure.
JP – We made a fair comparison between the two scenarios using similar
performance metrics. If you think that the analysis is biased,
can you make comments on the list.
Dimitri - Will provide comments on the mailing list.
Jean-Louis - This is interesting work, worth to be followed up.
================================================

================================================
20. CCAMP Work - Liaisons (Chairs, 20, 120/120)
   - Charter Update Discussion (cont.)
   - IEEE Related
   - SG15 Related
   - OIF Related
   Slides

Background reading
   - CCAMP Charter
   - CCAMP Correspondence

Will progress on the list.
Dimitri - Should MEF work still go through OIF or directly to MEF?
Adrian - Will try to find a way to talk with MEF
================================================