Last Modified: 2005-01-18
|Done||Submit IGMPv3/MLDv2 Interfactions with Routing Protocols as Informational|
|Done||Submit IGMP/MLD Proxying Specification as Proposed Standard|
|Done||Submit IGMP & MLD Snooping Considerations as Informational|
|Done||Submit SSM considerations for IGMPv3 & MLDv2 as Proposed Standard|
|Done||Submit Multicast Source Filtering API as Informational|
|Jul 04||Submit IGMPv3/MLDv2 MIB as Proposed Standard|
|Sep 04||Submit MRD as Proposed Standard|
|Sep 04||Submit MSNIP for IPv4 & IPv6 Specification as Proposed Standard|
|RFC3228||BCP||IANA Considerations for IGMP|
|RFC3590||PS||Source Address Selection for the Multicast Listener Discovery (MLD) Protocol|
|RFC3678||I||Socket Interface Extensions for Multicast Source Filters|
|RFC3810||Standard||Multicast Listener Discovery Version 2 (MLDv2) for IPv6|
Magma Meeting IETF 62
March 8, 2005
Isidor Kouvelas <firstname.lastname@example.org>
Brian Haberman <email@example.com>
Minutes for IETF 62 MAGMA Working Group meeting taken by Marshall Eubanks.
Brian Haberman - Agenda
Where do things stand ?
We are in good shape.
MRD is in AD follow-up
MIB has completed WG last call
one more rev to fix SMI issues
other documents are all in the RFC editor's queue
MRD Changes -
Jim Martin firstname.lastname@example.org
Some followup work after 04
Changes in 04
Russ Housley pointed out that we used IPSec AH.
Should switch to ESP. This is required by RFC2401bis, and AH
lacks a signature mechanism.
Added the potentiality of using a null encryption mechanism.
Key selection was unclear - now a symmetric key manually configured.
Thomas Narten - Several unclear references to advertisements.
NeighborDeadInterval is now 3 times the received Advertisement Interval.
Header field defined protocol to IGMP - so added ICMPv6 for IPv6
He thought that jitter was too large. They did not.
Bert Wijnen had some comments.
Alex Zinin asked for some clarifications.
So, those were the 03 to 04 changes. Thomas came back and was
unhappy with the checksum calculation, but they convinced him that
their calculation was OK, as it is consistent with the way
that other protocols are doing it.
He was unhappy about the jitter, so we changed it to what he suggested for 05 (but independently), with an explicit jitter interval.
We are waiting for further comments before submitting 05
Question : Isn't the jitter was the same ?
A. We narrowed the window on the jitter
Pekka Savola : There is a push in security for automatic key management.
Ross might through it back asking for that.
Dave Thaler : In the change from AH to ESP, you lose checking of
IP options and header fields. Given that the draft uses header fields
Someone pointed out that they could spoof a packet and send it to me by unicast. Maybe there should be language saying drop it.
Jim Martin : That is reasonable.
Pekka : It still needs to be a valid neighbor, otherwise the
ESP transport will fail. If this is a MUST, its ok, but if it says that you SHOULD use ESP, then there needs to be further checking.
Dave Thaler : It does not say MUST.
Bill Fenner : About the check sums. It wasn't clear to me - did you have to change it ?
Brian - for Julian Chesterfield - the MIB
Currently, after comments during last call, removed the RowStatus
ReadCreate functionality from the cache tables for security reasons.
We ran it through Smilint, and there were a few errors, which Julian will fix, and it will be ready for the IESG. If anyone has issues with the MIB, I would like to see them raised.
Brian - the Working Group Future
The two documents discussed here are the last ones we have. Of the
other two that were at one time considered, we are going to let MSNIP die a graceful or ungraceful death, and also the SSM MRD extension.
So, unless anyone has any issues that we should be raising, this will be the last WG meeting. Does anyone have any issues ?
Pekka : The working group will be closed ?
Brian : The working group will be closed when the document goes to the RFC editor's queue.
Scott Bradener : It's not when the document goes into the queue, but when they go out of the tree.
Dave Thaler : I would like to complement the working group for accomplishing what they set up to do. If there were going to be any changes in a new IGMP, that should be in a new working group.
Pekka : What has been debated on and off is extensions to anycast.
And there is also MAGMA security analysis document, which would be very useful.
Brian : There has been push back from the universe about how anycast should be and could be deployed.
Toerless : Why is anycast different ? Other groups do things that have had push back.
Brian : At one point we had discussed putting in anycast, but there didn't seem to be any interest and so it died.
Greg Daley : I think that we can get the MAGMA security document into the groups mailing list before it shuts down.