September 2008doc.: IEEE 802.11-08/1017r0

IEEE P802.11
Wireless LANs

Task Group U Meeting Minutes for
September 2008 Ad Hoc Meeting
Lihu‘e, Kaua‘i, Hawai‘i, United States
Date: 2008-09-03
Author(s):
Name / Affiliation / Address / Phone / email
Matthew Gast / Trapeze Networks / 5753 W. Las Positas Blvd
Pleasanton, CA94588USA / +1 925 474 2273 /

Wednesday, September 3, 2008

Acting Chair: Dave Stephenson

Recording secretary: Matthew Gast

Attendance

  • Gabor Bajko, Nokia
  • Necati Canpolat, Intel
  • Matthew Gast, Trapeze Networks
  • Dorothy Stanley, Aruba Networks
  • Dave Stephenson, Cisco

Call to order and agenda

Meeting called to order on Thursday, June 19, 2008 by Dave Stephenson at 9:06 am Hawaii Standard Time (HST). The chair then reviewed the following topics from the agenda:

  • The agenda is document number 11-08/1000r0
  • The chair read the IEEE patent policy
  • The membership had no questions on the policy
  • The chair requested information on essential patents, patent claims, and pending patent applications and called for letters of assurance. No response was made to the call
  • The chair also noted the affiliation FAQ, anti-trust FAQ, ethics code, IEEE 802.11 policies and procedures, and IEEE 802 policies and procedures

Comment Resolution

The current working comment resolution document is 11-08/0722r12.

  • CID 820
  • Dorothy Stanley: What is the advantage of separating these two features?
  • Dave Stephenson: An implementation must maintain both the counter and MIB table.
  • CID 901
  • Dorothy Stanley: It is also possible to re-order 7.4.7.12 so that multicast elements appear at the end
  • CID 13
  • DaveStephenson: There should not be a conflict between policy and what is provided. If a request would overridethe AP policy, it should be refused. Plus, conflicts between policy and requests should only be an issue for downstream flows because there is no such thing as an upstream TCLAS.
  • CID 46
  • Gabor Bajko: Why not just delete this element from the beacon and keep the list?
  • Dave Stephenson: The OUIs in the Beacon provide a shortcut. You can't say that because no providers have an OUI we shouldn't do this. Maybe this provides an incentive for them to get an OUI.
  • CID 690
  • Dave Stephenson: These objects are static information the carrier has based on location of a fixed AP.
  • Dorothy Stanley: Is the location data provided to a STA when it requests information?
  • Dave Stephenson: There is no facility for doing that in TGu.
  • Gabor Bajko: In the IETF, we want to define landmarks in civic location format and use them as a reference.
  • Dorothy Stanley: This doesn't propagate anything that isn't explicitly configured.
  • Dave Stephenson: These fields are an ASN.1 encoding of RFC 4119.
  • Resolution: Decline the comment by pointing out that it does not get transported over the air.
  • CID 87
  • Dorothy Stanley: Adding the personal area network type will cause confusion in the sponsor ballot stage.
  • CID 99
  • Dave Stephenson: Native GAS queries are responded to immediately; GAS multicast delivery allows for later delivery to a group of STAs.
  • Necati Canpolat: Multicast queries are useful because they allow stations to share the same query.
  • Dave Stephenson: GAS is not designed to be like NetBIOS. If it is important to have the information quickly, it is possible for a STA to snoop on the unicast responses to other STAs.
  • CID 811
  • Dave Stephenson: What if an AP is asked for an info ID that is not configured? If the info is not configured, the AP will never have it. If you have a list of info IDs, the AP can provide success indications for what it has.
  • Matthew: Fragmentation adds implementation overhead. The only way to keep this simple is to have a "you asked for too much data" packet. The definition of the info IDs right now makes it unlikely to exceed the MMPDU limit.
  • Gabor Bajko: Why are there different behaviors in the two types of GAS?
  • Dave Stephenson: Native queries are simple, and there are many operations that are defined as impossible.
  • CID 875
  • Matthew Gast: One potential use case is to restrict service based on location
  • Gabor Bajko: That is a poor security protocol
  • CID 613
  • Gabor Bajko: This is a good name.
  • CID 94
  • Dave Stephenson: These two items were disconnected to save Beacon bloat

The meeting recessed at 4:45 pm HST, with the comment resolution document uploaded as 11-08/0722r12.

Thursday, September 4, 2008

Chair: Stephen McCann

Recording secretary: Matthew Gast

Attendance

  • Gabor Bajko, Nokia
  • Necati Canpolat, Intel
  • Matthew Gast, Trapeze Networks
  • Stephen McCann, Roke Manor Research
  • Dorothy Stanley, Aruba Networks
  • Dave Stephenson, Cisco

Call to order and agenda

Meeting called to order on Thursday, June 19, 2008 by Stephen McCann at 9:10 am Hawaii Standard Time (HST).

Comment Resolution

  • CID 99
  • Necati Canpolat: Multicast is more efficient, so we need to have it for power saving. It is good to ensure that query responses can be picked up by everybody.
  • Dave Stephenson: The multicast address consumes space in the Beacon since it must be registered.
  • Matthew Gast: We could register a well-known multicast address with IEEE ANA, and then all stations could listen to GAS replies.
  • Dave Stephenson: How would we know whether native GAS would be multicast or unicast?
  • Necati Canpolat: We could allow a station to determine whether each info ID type is delivered multicast or unicast.
  • Matthew Gast: If multicast delivery is used, how do we ensure reliability?
  • Dave Stephenson: We would have to build the same response mechanism as in non-native GAS.
  • Gabor Bajko: In an airport when all devices are switched on, there will be lots of devices asking for the same information. How long will a device listen to the multicast address before it makes its own query and making the problem worse?
  • Dave Stephenson: It is hard to imagine why multicast queries would be used with today's protocols.

Straw poll (9:58 am): Should the task group retain the multicast delivery method in some form?

  • Results: 2 keep it – 1 remove the feature – 3 do not have strong opinions
  • Stephen McCann: This is a good feature but may be only useful in corner cases. It needs fixing.
  • Necati Canpolat: Simplicity is beauty
  • Stephen McCann: I will check out whether TGs will be using the multicast protocol.

CID 100

  • DaveStephenson: We should keep the protocol simple and leave it as-is.

CID 820

  • DaveStephenson: We can split off SSPN capabilities.

CID 83

  • DaveStephenson: Carriers are no longer attending and participating in the task group. We should retain authentication information (related to roaming) because the error codes are transmitted over the air. All of the rate limiting, permissions, and such can be done at higher layers. This would remove much of the MIB, clause 6, some of clause 11, and Annex K.

Presentation: (unnumbered document) IP address type, Gabor Bajko

  • Dave Stephenson: What is the point of this proposal?
  • Gabor: We need to assist network managers in the transition to IPv6

Presentation: (unnumbered document) Data/time format, Gabor Bajko

  • Dave Stephenson: This document needs a normative reference to ISO spec

Friday, September 5, 2008

Chair: Stephen McCann

Recording secretary: Matthew Gast

Attendance

  • Gabor Bajko, Nokia
  • Necati Canpolat, Intel
  • Matthew Gast, Trapeze Networks
  • Stephen McCann, Roke Manor Research
  • Dave Stephenson, Cisco

Call to order and agenda

Meeting called to order on Thursday, June 19, 2008 by Stephen McCann at 9:10 am Hawaii Standard Time (HST).

ES network discussion

  • Gabor Bajko: Requiring QoS end-to-end across the network a very very very very very bad idea according to NENA
  • Dave Stephenson: I disagree because the call needs to be unintelligible. NENA has jurisdiction only over North America, and there may be other opinions at other regulatory agencies.
  • Stephen McCann: We need QoS over the air for voip, but we can't say anything about the rest of the network
  • Gabor Bajko: Quality depends on the RSSI, which may be weak. Nena wants to deploy VoIP for emergency services by 2010, and if they require QoS it will slow down therollout.
  • Stephen McCann volunteered to write a liaison to NENA to clarify their position on QoS over the air and in the core network.
  • Gabor Bajko: There are two open comments on this point - CIDs 25 & 26
  • Dave Stephenson: Page 17 says that we are reporting what administrator has configured, not imposing requirements on external network
  • Stephen McCann: We need to make sure this text applies worldiwide. For example, Japan doesn't allow transmission of location on ES calls
  • Gabor Bajko: Make the last two bullet points depend on a MIB variable for regulatory rules
  • Dave Stephenson: NENA is not a regulatory domain
  • Stephen McCann: We don't want comments to depend on liaison response

Presentation: (unnumbered document) Native GAS query for Roaming Consortium and NAIrealm, Gabor Bajko

  • No comments on the presentation.

Presentation: 11-08/1028r0, GAS Comment Resolution, Dave Stephenson

  • After discussion, the ad hoc agreed to recommend removing GAS multicast delivery.

Agenda Planning

The group collaborated on a proposed agenda for the interim meeting in Kona.

The meeting adjourned at 3:28 pm.

Submissionpage 1John Doe, Somwhere Company