CCAMP Minutes For IETF 92
MONDAY, March 23, 2015
1520-1650 - Monday Afternoon Session II
Room: Oak
Presentation Start Time Duration Information
0 15:20 10
Title: Administrivia - WG Status - New charter and WG scope
Presenter: Chairs
0 15:30 5 Title: Reporting on WG drafts not being presented
- Draft-ietf-ccamp-ospf-availability-extensions-01
and draft-ccamp-rsvp-te-bandwdith-availability-01:
Amy Ye: Drafts adopted in August and just some
editorial changes since then. There are discussions ongoing regarding the
general applicability of the drafts.
-draft-ietf-ccamp-additional-sygnal-type-g709v3
and draft-ietf-ccamp-otn-signal-type-subregistry
Matt Hartley: Some changes since Honolulu based on the comments from Lou Berger and Adrian. There was the thought of extending the OTN subregistry and do a similar job on all the IANA registries in the scope of CCAMP but it has not been done yet. No outstanding issue on the drafts.
1 15:35 5
Title: Framework and Requirements for GMPLS based control of Flexi-grid DWDM networks
Draft: draft-ietf-ccamp-flexi-grid-fwk
Presenter: Oscar Gonzalez de Dios
Daniele Ceccarelli: As said before, there is IPR
polling for this draft. Please reply to the IPR polling. Secondly, since both
the chairs and Oscar are involved in this draft, Lou Berger will call for WG
consensus.(Thanks for that). If you have comments,
please don't wait until the LC. Any time is good to say anything.
2 15:40 10
Title: Flexi-grid protocols extensions - update and latest discussion report
Draft: draft-ietf-ccamp-flexible-grid-ospf-ext, draft-ietf-ccamp-flexible-grid-rsvp-te-ext, draft-ietf-ccamp-flexigrid-lambda-label
Presenter: Haomian Zheng
Lou Berger: Is there any reason for this extension
to be technology specific? Option 1, you are talking about using a generic
mechanism to start and seem it is a reasonable discussion, at least many people
think it is useful. The reason is we did generic before, why do tech specific
here? I don't have strong opinion, but worth asking this question.
Daniele Ceccarelli: the first thing that comes to
my mind is the differentiation between bandwidth and portion of spectrum.
Thinking of this, I would tend to say it is tech specific.
Lou Berger: This is a valuable point, and I also
said that on the list. We had the same issue with OO switch
and we have mems and mems
switch a portion of spectrum. We still at that time going to use bit per second
as coded by IEEE format, and we had done that before on g.709 drafts. We really
did that before in generic way. So, It really does not make sense to do
something tech specific when we could do generic and we don't need to worry
about next time. So think about it.
Rajan Rao: One of the parameters that impacts the Max LSP bandwidth bit per
second conversion is the modulation format, this was discussed on the mailing
list. Everybody thinks that it is related to other format, so people prefer to go
for technology specific. We can look at redefining Max LSP bandwidth as number
of the slots.
Fatai: Everyone agrees to introduce Max LSP Bandwidth but how to encode is
not decided.
3 15:50 5
Title: Link Management Protocol Extensions for Grid Property Negotiation
Draft: draft-ietf-ccamp-grid-property-lmp
Presenter: Xihua Fu
Daniele Ceccarelli: How many people think that this
draft is ready for LC? Not so many. How many think it is not ready and any
issues to be fixed? No open issues and take to the list.
Sergio Belotti: I have
concerns about the topic defined by the draft. This draft proposes to exchange
the capability of flexible and fixed grid. Since I suppose you don't have lots
of them, you have one or the other and you could probably configure before and
there is no meaning to exchange the information.
Fatai Zhang: So you mean that we could manual configure that, which is easier
than the exchange by the protocol?
Sergio Belotti: Yes, I
have doubts on the applicability.
Xihua Fu: In GMPLS many things can be manual configured by the management
plane and there is no conflict between the manual configuration and LMP
communication.
Fatai Zhang: It depends. For the dynamic info, we usually like to use control
plane to do that, for the static info, people may prefer to configure that.
4 15:55 10
Title: Information Model for Wavelength Switched Optical Networks (WSONs) with Impairments Validation
Draft: draft-ietf-ccamp-wson-iv-info
Presenter: Giovanni Martinelli
Daniele Ceccarelli: Please start drafting a text for the liaison, send it to the list for review/contribution.
5 16:05 10 Title: Information Encoding for WSON with Impairments Validation
Draft: draft-martinelli-ccamp-wson-iv-encode-05
Presenter: Giovanni
Martinelli
Daniele Ceccarelli: Liaisons does not
necessarily need to be tight to a WG draft. If you think it is easier/more
convenient to have a single liaison covering both this draft and the wson-iv-info one, please go ahead. Regarding the
experimental path I don’t have a strong opinion yet and would like to poll the
WG. We had a similar discussion for the WSON impairment free.
Lou Berger: Correct. If people are not sure to ship products based on a draft or there’s just one vendor, experimental is a nice way to go. Unfortunately experimental does not mean what it used to: the IESG decided to put a time limit on experimental drafts and this is moving away from them. My personal opinion is that this is a bad idea. If we had the WSON drafts as experimental they would have been available much earlier.
6 16:15 10
Title: A Yang Data Model for WSON Optical Networks
Draft: draft-lee-ccamp-wson-yang
Presenter: Young Lee
Daniele Ceccarelli: I think here
there is some work that could be considered technology agnostic and some that
is technology specific, e.g. connectivity matrix.
Young Lee: here we are considering
technology specific constraints. What defined in the upper layers is not
applicable here.
Lou Berger: comment on the
description of the first 2 bullets, they point at an old draft. Since then
there has been a lot of work on generic TE models (will be presented later this
week in TEAS). I think the right way to go is to augment those models.
Young Lee: When you say generic
model, what is generic is the label, but here we are speaking about WSON
devices and the label concept is contextual and we need to keep the label as a
lambda label. When we do info model we can generalize it, but here we are talking
about WSON in particular.
Igor Bryskin:
What you describe is a generic TE model but it is specific to a WDM layer, this
is what we have in mind by augmenting the models we’re developing with link and
node attributes. The connectivity matrix you describe applies also to OTN. I recommend
you to look into the existing model.
Young Lee: we don’t want to augment
more than necessary components. We’ll look for augmentation as much as
possible, when possible.
Gabriele Galimberti:
This is a good start. Do you plan to define also the amplifiers and so on?
Young Lee: It will be the next step.
Gabriele Galimberti: You said you want to augment existing models, for the interfaces I would suggest to consider G.7223 which is a good starting point for DWDM interfaces.
7 16:25 9 Title: Extension to the LMP for DWDM Optical Line Systems to manage the application code of optical interface parameters
Draft: draft-dharinigert-ccamp-g-698-2-lmp-09
Presenter: Gert Grammel
Haomian Zheng: Is the power monitoring related to the
frequency or is the same for all frequencies?
Gert Grammel: It’s not
related to the wavelength, but since a black link is a wavelength, we could say
that the power monitoring is related to such wavelength.
Haomian Zheng: I’m asking because if we make it mandatory
it is going to be very expensive, mostly for higher rate interfaces.
Gert Grammel: Agree, that’s why it is not a mandatory feature. In addition to
that not all the equipments would be able to monitor
the power. Anyway this is about protocols; we don’t specify how the power is
monitored but how the info is carried.
Daniele
Ceccarelli: When you say that LMP is used nor for signaling or for routing, you
mean that you’re not dealing with routed paths, in the sense that they are
fixed paths?
Gert Grammel: A link is black
when it is lighted up, in the sense that once you’ve lighted it up you have a number of parameters that characterized it, not
before. In addition to that we’re not
asking to setup a wavelength, what we do is informing the guy that is setting
up a wavelength what are the options; so the answer to your question is yes.
Fatai Zhang: (comment on slide 6): You are saying that
you have one end of the link that supports application codes A and the other
supporting application codes A, B and C and you’re exchanging this kind of
capabilities, right? This is a quite static
information, wouldn’t it be enough to configure the transponder to work with
application code A?
Gert Grammel: It is not
either discovered or configured. This is an additional option.
8 16:34 8
Title: An SNMP MIB extension to RFC3591 to manage optical interface parameters of "G.698.2 single channel" in DWDM applications
Draft: draft-galikunze-ccamp-g-698-2-snmp-mib-09
Presenter: Ruediger Kunze
Daniele Ceccarelli: Reading the
minutes of the last meeting I see that there was the proposal to remove the
power as a parameter. I see the power is still there, is there any particular
reason not for removing it? Are you planning to put it into the vendor specific
parameters or whatever?
Ruediger Kunze: In DT we have different departments for the provisioning in the different layers. In order to give the optical guys the opportunity to configure the optical interfaces in an easy way we decided to allow to set the power here. Now we have a YANG draft for configuration, so no need to keep the power config in SNMP. For DT is ok to remove the power setting from SNMP but not the info getting.
9 16:42 8 Title: A YANG model to manage the optical interface parameters of "G.698.2 single channel" in DWDM applications
Draft: draft-dharini-netmod-g-698-2-yang-02
Presenter: Gabriele Galimberti Galimberti
Xian Zhang: The 3 drafts are dealing
with the same parameters, are they completing each other or competing with each
other? At least the MIB and YANG ones seem to be competing.
Gabriele Galimberti:
It’s just a
matter of time. At the time being SNMP is the most used protocol. Once YANG
will take place, probably the SNMP MIB draft could be “frozen”.
Xian Zhang: Application codes should
be standard parameters, only get from the network, not configured.
Gabriele Galimberti: Correct. The application codes are get only. What can be configured are parameters like power and frequency.
Adjourn 16:50