2.5.1 General Switch Management Protocol (gsmp)

NOTE: This charter is a snapshot of the 49th IETF Meeting in San Diego, California. It may now be out-of-date. Last Modified: 30-Oct-00


Avri Doria <avri@nortelnetworks.com>
Kenneth Sundell <ksundell@nortelnetworks.com>

Routing Area Director(s):

David Oran <oran@cisco.com>
Rob Coltun <rcoltun@redback.com>

Routing Area Advisor:

Rob Coltun <rcoltun@redback.com>

Mailing Lists:

General Discussion:gsmp@revnetworks.com
To Subscribe: gsmp-request@revnetworks.com
In Body: subscribe (unsubscribe)
Archive: ftp://www.revnetworks.com/pub/mailing-lists/gsmp-archive

Description of Working Group:

The General Switch Management Protocol(GSMP)has been available to the IETF community for several years as an informational RFC. As several vendors are beginning to deploy GSMP, there is now a need to standardise the protocol.

GSMP provides an interface that allows a router to control a label switch. The GSMP interface provides the commands necessary for: switch configuration control and reporting, port management, connection control, QoS and traffic engineering control and the reporting of statistics and asynchronous events.

The working group is responsible for standardising the GSMP protocol itself, as well as for defining an encapsulation for running GSMP over IP. The current GSMP V2 set of message types will be expanded to include support for various types of label switches including ATM and Frame Relay. The working group will refine the support for QoS and traffic engineering in GSMP based on the diffserv and intserv architectures and will co-ordinate the work with these WGs where appropriate.

Finally, since GSMP can be used as an adjunct to the MPLS protocol, the group will co-ordinate with the MPLS group to make sure that the primitives required by a MPLS controller are available in the protocol.

Goals and Milestones:

Apr 99


Define an IP encapsulation for GSMP messages

Apr 99


Extend GSMP message types to allow support of non-ATM Label Switch devices

Aug 99


Refine the support for QoS models in GSMP

Nov 99


Define an approved GSMP MIB

Nov 99


Define approved mechanisms for security and authentication in GSMP

Dec 99


Produce a version of the GSMP which has the consensus required for entry onto the standards track.


No Request For Comments

Current Meeting Report

The GSMP working group held a meeting at IETF49, Thursday, December 14, 1300-1400
The meeting was chaired by Kenneth Sundell and Avri Doria, and notes were taken by Hans Sjostrand.

1. Agenda bashing

The agenda was presented.

The 2nd last call closed without any substantive comments on the three documents. The fourth document, the applicability document, didn't have any substantive comments from the first last call.

Two proto work items will be presented. The TDM label doc and the mibs and the pibs that where approved for proto work items awaiting an approved charter update.

Then a conversation on the charter update. It was mentioned that the gsmp WG is considered for the new sub-ip area.

The chair asked the working group if they wanted to add, change or reorder any item. There were no comments.

2. Last call comments

- draft-ietf-gsmp-08.txt

Kenneth Sundell presented the updates of the gsmp spec. The last call went out after the Pittsburgh meeting, second last call a week before this meeting. A detailed list of comments and responses/changes has previously been posted to the list. Lot of cleanup work, because of history with four different editors, to get the spec more consistent. Thanks for the help with comments.

Some major changes, TDM labels where taken out, folded into another document that will be presented later this meeting. Also, all labels are now encoded as TLVs, the short labels where removed.

Minor changes, e.g. the PTYPE is updated to allow the controller to assign partitions and multiple FEC element enhanced the FEC Label type.

Rev -08 passed without comments, next step is IESG last call. What the chairs have done is to let the AD know that the WG are about to submit 4 specs to IESG, they ok'd it.

- draft-ietf-gsmp-mib-03.txt

Hans Sjostrand presented the mib updates.
The object structure is easy and symmetrical.
The notifications where unchanged from the previous version.
The changes where posted to the list and are also listed in the I-D.
Virtual was removed from all names, the mib was reviewed by a mib expert so the mib got plenty of snmp updated, e.g. proper internet addressing, TCs added etc.
Also partition type was added and is corresponding to the one in the base spec.

There where one comment made at the last call, to clarify gsmpSessionDiscontinuityTime object description.
Also an editorial comment to get rid of the funky ¡ characters.
Also a contact change, Hans is moving from Ericsson to IPunplugged.
An update is early next year.

Comment from the AD that there is no big change so there is nothing for the WG to consider, the AD review will start and we just notify them when the updated version is out.

It was asked whether it compiles. Yes, it compiles using several well-known compilers including libsmi. There is a problem with SMICng because of a bug in that compiler in the handling of access-to-notify so that needs to be handled with a compile option.

- draft-ietf-gsmp-encaps-03.txt

Joachim Buerkle presented the encaps updates.
The last call comments were fixed, the majority of changes were due to the comments from Tom, which was mostly editorial in nature.

No comments on the 2nd last call. It's ready to go to IESG on rev 03.

Tom Worster wanted to know the expected time before the IESG has finished their review. Rob Coltun answered that the IESG is prepared, but it's a review process where they could be deferred for a while for updates etc. and it is unpredictable. Also, given the degree of controversy, it might be long.

3. G-MPLS support, draft status

None of these I-Ds where submitted, this is just a status information.

- TDM labels new draft

Jonathan Sadler presented it. G-MPLS label formats has been solicited for SONET/SDH, label formats for PDH have been proposed.

Should be common constructs as for G-MPLS. There are some changes, the biggest change is how fractionated services are done.
Label list vs. bit field encodings. New services determination e.g. whole port switching, needing a port label.

The remaining work includes capability determination and traffic characteristics. Also inclusion of service layer constraints.

Assistance and help is more than appreciated.

What's not included?

Packets label format and characteristics. Optical labels which are subject to working group re-charter. Some formats were submitted informally last meeting. Which way we go has to be discussed within the WG.

- MIB/PIB for switch partitioning.

Todd Anderson gave a status report. A first version was presented at the last ietf. To assign virtual switches based on partitions. The pib was converted to a mib. A virtual link was added to allow connection between partitions within the same switch. Also, the mib was compiled and the problems due to that were fixed. A new rev is ready and will be sent to the list next month for comments. After these comments are resolved the MIB will be converted to a PIB aswell.

Some open issues. GSMP control is configured for partioned switches, maybe other controllers wanted to be partitioned, e.g. megaco controllers or TDM switches. He requested for volunteers to start working on these issues..

Avri noted that the mib/pib is focused to pertain to the stuff that are going out now, to get the resources tied to static partitions. There have been requests from service providers that the they want dynamically assigned partition resources without reboot of the switch. That could be part of the recharter. Then we need a notion of the switch model that we are partitioning. When we get more general we have to be more rigorous in the models.
Working group input on the mib is needed.

4. Charter Status update.
The proposed charter has been submitted to the IESG. This charter has reached consensus on the gsmp mailing list..

The new charter is adding support for optical and TDM switching. Defining mechanisms for switch partitions and mechanisms for control of IP packet switching devices. This is up for negotiation if that work should be made at all, or where it belongs. It's switch management support protocols kind of work but definitely widening the scope of the WG.

Regarding support for optical switching, the WG will add capabilities to the GSMP spec in cooperation with the IPO WG or whatever WG that there might be concerned with optics.

There where a comment from Rob Coltun that we have to gauge the industry support, there is no question that the tech is cool, but has it been developed and deployed?

Avri responded that there are two types in this industry, first the service providers wanting the products, second come the vendors having the products. Within the context of MSF (Multiservice Switching Forum) which is about half service providers and half vendors, there is a strong interest from service providers, especially to be able to get new services without forklifting their network.

Part of the discussion was regarding optical and TDM that there is a very strong interest to se it, GSMP where too late for ATM. Avri mentioned that the company she works for is currently designing this capability into some of its equipment. There is also conversation about how much different vendors are willing to talk about it in public.

Fred Baker asked; How do we know that there are vendors and that there are customers. How many vendors are supporting this and are there any networks supporting it, has it been deployed anywhere?

Avri argued that as soon as it's proposed standard, we need to have the protocol first and sometimes later we have the networks.

There is an interop event being planned for q3 2001 with four vendors that have announced interest.

Jerry Sydir of CPLANE announced that they had a product for GSMP controllers. Also there have been lots of extensions by the service provides themselves. They want to do it themselves to get the control of the services in the network.

Olle Pers of Telia announced that as a service provider they would like to have these interfaces, especially on optical equipment. And they want to have an open standard and not vendor proprietary.

There was opposition when the TDM was removed, that work was added because the companies pushed it to the WG. There is a draft in progress. That draft looks in to the traditional cross-connect perspective. It could have a more G-MPLS like perspective.

Avri noted that the preliminary interest is mainly technological and getting the protocol out. There are proprietary products deployed. One very major company has deployed the control entity with big success. It's not that hard to extrapolate from that.

Any more information could be sent to the IESG.

Tom Worster noted that there is a little of the chicken and egg problem. There has been a fair amount of interest and people are doing proprietary stuff. We want to get the standard out as quickly as possible.

The IESG desired the carriers saying that the are ready deploy this, that it is exactly the right way. We missed the first wave with ATM but maybe we are ready for the second wave.

Whether we could get testimonials from carriers that they want this (more than Telia who spoke up here now) we should get it.

The milestones where presented by Avri. We slipped one year last time but this time we have a good chance to be in time.

Abel Weinrib made the comment that you skipped over the control of IP bullet, for this WG it feels like a sidetrack. For example, it's coming as third item in order on the charter list.

Avri responded that there are a lot of interest in it here, but it shows up as third because it's controversy in the discussion with the ADs. It's not sure that GSMP is the right vehicle for this. Maybe a PIB is a better vehicle for this. It's really depends on where the IESG expands the charter.

Abel Weinrib noted that FORCES wants the people over to the forces bof to do the work there.

5. Intrerop event

MSF is thinking on sponsoring an interop event. There is also a prospective corporate sponsor (that is not Nortel). It will be a phase one testing, there will be a test on the specification and not on specific implementation. There will be no report about how the equipment performed or not, the discussion will be around how the specs perform.

Meeting adjourned by Avri Doria.


GSMP Packet Encapsulation for ATM, Ethernet and TCP
Definitions of Managed Objects for the General Switch Management Protocol
TDM Labels
Proposed GSMP Charter