Decisions
- Probably last face-to-face WG meeting. Submit drafts to the IESG, keep the mailing list running at least until the drafts have passed the IESG review.
Action Points
Open action points from IETF-57
- IETF57-AP1: Re-charter
-
Status: Closed. (Dec 5th)
Later decision: The WG will skip the re-charter. The re-chartering would have been only technical in nature (not touching the real content of the WG, and the WG is closing.
- IETF57-AP8: WG LC on draft-ietf-send-ndopts
-
Status: Delayed. (Dec 5th)
New target date: December 2003.
New action points from IETF-58
- IETF58-AP9: Advance draft-ietf-send-psreq at the IESG
-
Owner: James Kempf
Target date: November 2003
Status: Pending IESG. (Dec 18th)
Document scheduled for IESG meeting on Dec 18th
- IETF58-AP10: Submit draft-ietf-send-cga to IESG
-
Owner: Pekka Nikander
Target date: November 2003
Status: Pending editor. (Dec 5th)
Editor making last minor modifications. Draft expected to be submitted on week ending on Dec 12th.
- IETF58-AP11: Clarify measurements from implementation report
-
Owner: James Kempf
Target date: November 2003
Status: Closed. (Nov 28th)
Details sent to the mailing list.
- IETF58-AP12: Discuss CGA address selection at mailing list
-
Owner: Pekka Nikander
Target date: November 2003
Status: Closed. (Nov 31th)
Discussion ongoing.
- IETF58-AP13: Cover different functions of Redirect
-
Owner: Jari Arkko
Target date: December 2003
Status: Open.
- IETF58-AP14: Add text on unsolicited ND to NDOPTs
-
Owner: Jari Arkko
Target date: December 2003
Status: Open.
- IETF58-AP15: Remove excess text from NDOPTs
-
Owner: Jari Arkko
Target date: December 2003
Status: Open.
- IETF58-AP16: WG last call & board review on NDOPTs
-
Owner: Chairs
Target date: December 2003
Status: Pending AP13-15.
Milestones
Review of open milestones
- Dec 03
- Submit draft-ietf-send-cga-xx.txt to IESG for approval.
- Expected to be completed on week ending on Dec 12th.
- Dec 03
- Submit draft-ietf-send-ipsec-xx.txt to IESG for approval.
- May be delayed until January 2004.
Charter review
Technically, the WG should be rechartered to reflect the fact that the WG will produce an ND OPTions based on approach instead of an IPsec based approach. However, it was decided that the rechartering will not be done since the WG is closing to end anyway.
Discussion
- Pasi Eronen:
- I reviewed the CGA draft and even checked the hashes in the examples, and the draft seems to be OK.
- James Kempf:
- Thanks, these kinds of reviews are important, especially ones that go deep enough.
- Greg Daley:
- Did Jon Wood implement the DCS/DCA response delay? In other words, did the measured number include the DCS/DCA random delay?
- James Kempf:
- No, I don't think so. I will post a definite answer to the list.
- Margaret Wasserman:
- Question. HOW do we configure to use the secure address as the source address? All the other address selection rules in RFC3484 check the prefix of the address. How can the configuration table in RFC3484 dtermine if the address is a CGA address? Or do we always want to prefer the secure address?
- Christian Huitema:
- We already do this with RFC3014 privacy addresses. We might not want to prefer a CGA address, we might want to use some other address, a throw away address which we use for privacy. We may also want to use a different key for privacy addresses.
- Samita Chakrabarti:
- Answering to Margaret's question, in terms of source addresses it is implementation specific how to determine whether to use a CGA address or not. The local node will know if a given source address is a CGA or not. The MIPv6 API contains switching of addresses on/off.
- Jari:
- For care-of addresses, one might prefer, but why CGA?
- Samita:
- I am not sure, I am probably not the best person to answer this question. But please specify what the default mode is to be, to use CGA addresses or not? For the question of whether a new API should be created or not, please consult the IPv6 WG.
- [Clarification by Samita, send afterwards:
- draft-chakrabarti-ipv6-addrselect-api-02.txt already has a mechanism to choose CGA and NON-CGA source addresses by the application using socket api. The SEND wg needs to decide whether such API is needed for applications running on a SEND-node. Please send the decision on this to the IPv6 wg alias - based on that we will update the address selection API draft (which actually goes along with RFC3484).]
- Tony Hain:
- The new public keys will have to registered somewhere, right?
- Jari Arkko:
- We do not need any infrastructure to store the keys.
- James:
- These keys aren't being used elsewhere. Only in RS/ND.
- Tony Hain:
- If CGAs become the preferred addresses, we need to ensure that there is a place to make the keys available, since people will want to use the keys elsewhere. So, don't go there right now.
- James Kempf:
- We need to take this to the list.
- Greg Daley:
- I have not seen any hosts that send redirects. However, if there indeed are such cases, maybe we have to have a look into this.
- James Kempf:
- Lets take this, too, to the list.
- Pasi Eronen:
- I did the review of the NDopts document, too, and most of the issues are editorial in nature. For example, the ndopts document says the modifier is 16 bits and the CGA document states that it is 16 octets (or was it the other way around?). However, I have an issue with the timestamp algorithm. I am not sure whether it works with 1 second resolution for timestamp, even with less than 1 second timestamp. Allow things to be off by 1 second. But I've sent this comment to the list already. There seems to be also some text missing about unsolicited ND. If the node has some reason to update things, it can send them.
- Jari Arkko:
- Yes, we need to handle these issues. Thank you for bringing them up.
- Pasi Eronen:
- There is also lots of redundant text.
- Jari Arkko:
- Yes, the document needs some serious editing.
- James Kempf:
- The goal right now is to get the document ready, and announce another last call before the end of November, if possible. We will also send the document to the review board. Apparently, it will take some time to process the review comments, but the goal is to get the document ready for IESG before the end of the year. Then it usually takes 6-9 months to get through the IESG and the RFC Editor.
- Margaret Wasserman:
- We need to try to do better than 9 months for getting this through.
- James Kempf:
- The idea is to shut down the working group of the IESG review, but keep the mailing list running anyway.
- Margeret Wasserman:
- Maybe we need to keep the group on a hiatus if we need to go to draft standard. But that is just a process issue.
- James Kempf:
- I like things to start, cruise, and reach end. Hence I would prefer to close down the working group and run later another bof, if needed.