IETF/IEEE 802 Liaison Report
Dorothy Stanley’s March and May reports on the IEEE 802.11 liaison to IETF are available here:
David Harrington has requested that the IEEE 802.1 MIB transfer document be published. The document is currently in Expert Review and is available for inspection here:
With respect to the IEEE 802 side of the transfer, the IEEE Legal Counsel sent the following advice to IEEE 802.1 on April 12, 2006:
“Today, received signed permission letter from Mr. David Harrington, granting permission to use IETF RFC 4363 (Specifically P-BRIDGE-MIB and Q-BRIDGE MIB modules0 and IETF RFC 4318 (specifically RSTP-MIB modules). We need not await to receive any further permission grants with respect to the stated modules, as written permission from one co-author is sufficient under the law.”
>From the IEEE 802.1 perspective, this seems to open the way to start work on IEEE 802.1 MIB modules derived from the IETF Bridge MIB modules.
Dinesh Mohan has declined appointment as IETF liaison to IEEE 802.1. The position therefore remains open.
On May 15, 2006 the IEEE 802.11 ad-hoc group voted to approve a draft review of the PANA Framework document:
The document was brought to a vote of IEEE 802.11 on Wednesday May 17, 2006 and passed.
IEEE 802.16 has sent the following liaison letter to ITU-T relating to operational requirements for IP operation over Mobile Systems:
On February 10, the IEEE 802.16 liaison to IETF, Jeff Mandin, sent a reply to an inquiry originally submitted by Margaret Wasserman. The email is enclosed below.
Within 802.16 work is now underway on 802.16k (http://standards/ieee.org/board/nex/projects/802-16k.pdf ) which will amend 802.1D so as to clarify how transparent bridging should operate over 802.16. The initial 802.16k call for contributions resulted in a single proposal (
http://wirelessman.org/docs/06/S80216-06_001.pdf ) which was then adopted as the wg draft for letter ballot.
This work appears most likely be useful for backhaul and other non-mobile deployments.
At a recent meeting the General Packet Convergence Sublayer (GPCS) was accepted in 802.16g. See:
However, the WiMAX Forum Mobility Task Group has determined not to use GPCS, so that its influence is not clear. 802.16-2004 included a “no Convergence Sublayer” option. Cor1 then deleted it. GPCS in effect restores the “no CS” option together with some other functions.
IEEE 802.16f OID Assignment Issue
In the liaison letters below, IEEE 802.16 requested assistance in review of the 802.16f MIB:
As part of the work on IEEE 802.16f, 802.16 utilized an ifType number once assigned to a now defunct company for use by its proprietary products:
propBWAp2Mp (184), — PropBroadbandWirelessAccesspt2multipt
Unfortunately, 802.16 assumed wrongly that transmission 184 had been assigned to them and developed the 802.16f MIB under this OID, using an assignment under the IETF managed mib-2 subtree without asking IANA for an assignment. Currently no SDO other than the IETF owns a node under transmission.
The 802.16f MIB will be amended/superceded by the 802.16i project, currently in its early phases. However, 802.16f has been approved, and implementations are underway.
The current proposal under discussion is for 802.16 is to register a new ifType with the IANA and use this ifType in the 802.16i MIB. The 802.16i MIB will eventually supersede the 802.16f MIB, and therefore 802.16i can say that the new ifType number will be used going forward, but the old (184) may be used by old equipment. In addition, it is proposed to ‘legalize’ the usage of transmission 184 and so that 802.16 can continue to develop their MIB under this OID.
16ng Conference Call
A conference call was held on Friday, April 21, 2006 relating to IP over 802.16. This call involved members of the IESG, IAB and participants in IEEE 802.16 and the WiMAX Forum. Notes from the call are enclosed below.
Date: Fri, 10 Feb 2006 17:01:19 +0200 From: Jeff Mandin <firstname.lastname@example.org> To: Bernard Aboba <email@example.com> Subject: Re: Question for IEEE 802.1(6)? Cc: firstname.lastname@example.org, email@example.com, firstname.lastname@example.org, "Roger B. Marks" <email@example.com>, Phillip Barber <firstname.lastname@example.org>, "Johnston, Dj" <email@example.com>, firstname.lastname@example.org
802.16-2004, 16e, and now .16g have for the most part endeavoured to remain open to multiple schemes for IP transport. 802.16 Working Group discussions have generally assumed that different schemes are appropriate for particular deployment environments (eg. PPPoE is viewed by some as an appropriate approach when .16 is used for DSL replacement, but perhaps not in other environments such as mobile) and that these determinations would be made by vendors or in other standards organizations.
As you suggest, technical cooperation between 802.16, IETF, and the 802 Standing Architecture Committee on this issue could be beneficial. Such cooperation would best be initiated by the IESG sending a formal liaison letter to 802.16 so that the WG membership and 802 Architecture could be made aware of the issue.
Please note that there is a proposed PAR for .16 that might have some bearing on the work that you are describing. If approved, 802.16k ( http://ieee802.org/16/docs/06/80216-06_003.pdf ) will amend 802.1D so as to clarify how transparent bridging should operate over 802.16 (some more background is at http://ieee802.org/16/arc/802-16list2/msg03127.html ). This project would seem to clarify at least one method of IP transport (since mechanisms for IP transport over 802.1D would seem to apply regardless of whether 802.16 is used for part or all of bridged network). Of course other methods of IP transport are still possible.
Please let me know if there are any timeframe issues here, or if I can provide any additional information or assistance.
IEEE 802.16 liaison to the IETF
On 1/26/06, Bernard Aboba <email@example.com> wrote:
The IESG is now considering the chartering of work on IP over IEEE 802.16.
Is this something that IEEE 802 has previously given guidance on, or is it possible that it will wish to give guidance on the subject in the future? I am curious whether it would be possible to get feedback on this subject, potentially from within IEEE 802.16 or even at the IEEE 802 architecture level.
From: Margaret Wasserman [mailto:firstname.lastname@example.org] Sent: Wed 1/25/2006 5:27 PM To: Bernard Aboba Subject: FW: 16NG BoF Schedule Request
FYI — the 16ng folks are working on a problem statement and asking for a second BOF in Dallas.
They think that there is IETF consensus to work on this issue, but I am still uncertain why the Ethernet compatibility mode is inadequate for their needs. Their problem statement is supposed to answer that…
In the meantime, I am wondering how the IEEE feels about us doing this sort of work. Has it been discussed there? Do they have any particular opinion about how it would be best to run IP over 802.16?
Notes from the “16ng” call between IETF and WiMAX Forum on Friday, Apr 21st, 2006
Prakash Iyer – NWG chair at WiMAX Forum
Basavaraj Patil – IPv6 subgroup chair within NWG/WiMAX Forum
Soohong Daniel Park – 16ng BOF chair at IETF
Gabriel Montenegro – 16ng BOF chair at IETF
Jari Arkko – INT AD at IETF
Bernard Aboba – IAB member at IETF
2. State of the 16ng proposal at the IETF
Relatively positive feedback on WG formation in Dallas, including IESG/IAB participant’s comments.
The WG proposal is currently in IESG/IAB/BOF chair internal review. The topic is on the IESG April 27 telechat agenda, to decide whether it should advance to “external review” where it will be distributed within the IETF as well as externally. After external review another decision will be needed to actually approve the WG.
The main remaining issues appear to be clarification of responsibilities in this space between different standardization organizations, timing, and whether multiple different standards track solutions for the same functionality are appropriate (e.g., IPCS and Ethernet). One suggested approach to solve the last issue is to consider one of the CS layers as a standard and make others informational.
3. Overview of WiMAX Forum activities in defining “IP over 802.16/WiMAX”
WiMAX Forum is working on a profile of 802.16 to narrow its scope for interoperability purposes. This profile includes both radio and higher level issues. The IPCS has been defined as mandatory and Ethernet CS as optional as a part of this profiling work. There are some deployment scenarios where Ethernet CS would be used (DSL-type networks and multi-host scenarios, where there is more than one host in a CPE domain), but it is expected that most deployments would be using IPCS.
The profiling work is almost complete.
NWG started work on IPv6 in late 2005. When the NWG works in the IP parts of its network architecture, it relies on IETF RFCs (or in some cases WG Internet Drafts).
NWG wants to complete work by end of Q3, and Q4 is only used for editorial clarifications. The real deadline for IETF output (approved specifications, not necessarily RFCs) is therefore at the end of Q3, 2006. The specifications from WiMAX Forum will go out at the end of the year, however.
Basavaraj chairs a group that focuses on IPv6 aspects. Specifications come out in stages as in 3GPP, i.e., stage 2 contains architecture and stage 3 detailed protocol specifications. The work is currently in stage 3, but there is a need for an occasional return to stage 2 to correct errors.
The documents are accessible only to NWG members, but there is a recent agreement in the forum to make the latest drafts publicly accessible. This has not been announced officially yet, but some of the documents are already accessible from the public web site. The rationale for making documents accessible is the forum’s desire to work with other bodies, including co-operation in the 3GPP SAE/LTE work. Documents should be officially accessible in few weeks.
The structure of the specifications is broken down by function, address assignment, ARP, neighbor discovery, mobility, etc. The 802.16 specifications do not include all details about these for all CS layers.
Bernard: Just for my clarification, is v4/v6 IPCS now being made public? Basavaraj: IPv4 over IPCS is not really defined at the moment, but IPv6 IPCS is in the document that is being made public.
Bernard: Which ones of these specs are done and which are not, and where IETF can complement them? Prakash: We would appreciate recommendations on how IETF thinks ND/ARP etc. should be implemented. The rest, such as localized mobility management are not as urgent.
Jari: The basic question is who owns the specs. What are your expectations? Prakash: Our preferred approach is that there is an IETF proposed standard that can be used by reference. But the main problem is timing. We can use the specifications if they are available at the right time. There are, however, also future releases so we can put the references into them if things take longer than expected. Text could be added to current version to point to ongoing work.
Jari: Can we copy your specifications as a starting point? Prakash: Yes.
Gabriel: Timing – are the current dates in the charter proposal sufficient, or are they way off? Basavaraj thinks that they are doable but its tight.
Jari points out that its important to realize where the effort mainly lies in. We expect the main body of the work in the proposed WG to come from the 802.16, WiMAX, and WiBro people. So whether its doable from the IETF side depends largely on these people. We are prepared from the IETF to do it, the process, IESG approval and so on should not set the limit — the limit is really in whether the WG participants can create high quality specifications in the given time. One thing that is harder to estimate is whether we will run into technical problems, such as if IPv6 experts uncover issues that so far have not been discovered by others in this space.
Bernard: Do the current Internet Drafts reflect material in WiMAX Forum specifications? Basavaraj says that this is not the case yet, though the IPv6 CS draft may be a bit closer to this than others. Gabriel thinks they have to start from scratch.
Jari emphasizes also the importance of not standing still — drafts should be made available soon and technical work should progress, even if we are still discussing the WG approval. Daniel indicates that a draft is being prepared.
WiMAX Forum has not worked on the Ethernet versions of the specifications. It is expected that the Ethernet documents are independent from the others and can be progress in their own timescale.
5. Liason issues
A formal liason relationship is missing between the IETF and the WiMAX Forum. Such a relationship has been discussed before although at that time the discussion was not focused on the 16ng effort but rather a number of other working groups (e.g., RADEXT).
The IETF’s focus on any joint work is on direct, practical work at the WG level, such as the same individuals participating in different bodies or what this conference call attempts to find out about the expectations in the 16ng case. Liason arrangement can be used to support this practical level, but the practical level is most important.
Specification ownership: Ideally, IETF would have a proposed standard on “IPV6 over IPv6 CS”, and WiMAX Forum would refer to it. *If* timing doesn’t work, then start and seed the IETF docs, and when they are proposed standard, use them as references in a subsequent revision.
Timing: WiMAX comments resolution: late August, Wibro comments due in July based on their experience, final decision point is IESG approval. WiMAX needs stability by end of 3rd quarter. Proposed 16ng milestone deadlines are suitable, but can not slip if a reference to the first release of the WiMAX specifications is wanted. IETF is willing to make there are no process hoops to take too much time, but WG participants need to deliver high-quality specifications soon in order for the schedule to be possible.
Liaisons: Possible if we feel it would be helpful. But liaison is not necessary to get work done and track dependencies.