jari.arkko at piuha.net
Thu Jul 10 17:00:49 UTC 2014
I thought it would be useful to start discussion of some of the topics that the group needs to deal with. I’m sending you some thoughts on the timeline of events looking forward. This is a rough draft and for discussion purposes only. Please comment!
What is described below reads as one timeline, but it may be important to understand that we actually have multiple parallel efforts: protocol parameters, addresses, and names. These efforts have several points of linkage, but allowing parallel operation is probably a good idea. As a result, the timeline may run at slightly different speeds for the different parts.
The NTIA has said that the September 2015 deadline is not firm, and that they could extend contracts and prolong the process beyond that. I do believe, however, that there are several reasons why getting done by then is at least useful if not even necessary. First of all, I think we all need a goal. And open-ended timelines lead to open-ended discussions. Second, people, administrations, political climate, etc. may change. So I think we should work towards September 30, 2015. But of course, we also want a good proposal to go forward.
In thinking about what is needed before completion, I think see at least the following tasks:
1. Communities: Communities coming up with plans.
2. Coordination and alignment. This includes possible iteration and even going back to the communities. It also includes, possibly, some acceptance phase with the global community for the assembled plan. For instance, the assembled complete plan should probably be something that the customers of IANA are happy with. In other words, coordination can be lead by the coordination group, but it likely includes a lot of work also in the communities.
3. Test run. We may want to show off that the system can actually run in the proposed way for a while.
4. NTIA evaluation and approval process.
The community processes are in my opinion the most important ones, and should involve most of the work. In addition, they tend to run for long times. For instance, the simplest IETF last call would be a month already. We also want to give time to the NTIA to do what they need to do. And we should not optimistically assume that there is no need to adjust and iterate over the initially submitted plans. Which may include some additional time confirming with various communities that changes are OK.
I’m a little unclear how much testing/demonstration we need. Arguably, the IETF probably runs the system it needs already. But there are other parts where some re-organisation and re-thinking is needed. Not really clear to me how much confidence building would be needed for those parts. Does anyone have thoughts on this?
In general, I think we should run this process as the much ourselves as possible, and have everything (including criteria fulfilment) very clearly spelled out for the NTIA, rather leaving any substantial work for them. But they’ll still need some time to process. And we might again need to come back and change something.
I guess what I’m coming to is trying to get going pretty early, so that we can iterate in later stages.
The tasks are obviously overlapping. For instance, the coordination group should do coordination from day 1, by understanding what the communities are doing and pointing out if they are going into conflicting directions or if there are parts that are not being addressed.
So here is one possible overall timeline:
Step 1: Communities' work - ready by Dec 30, 2014 (6 months)
Step 2: Coordination and alignment, including iteration - ready by March 30, 2015 (+3 months)
Step 3: Acceptance and communication - ready by May 30, 2015 (+2 months)
Step 4: Testing - ready by July 30, 2015 (+ 2 months)
Step 5: NTIA evaluation and acceptance - ready by September 30, 2015 (+2 months)
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 842 bytes
Desc: Message signed with OpenPGP using GPGMail
More information about the Internal-cg