Re: issue 11 -- window size
"Mohan Parthasarathy" <[email protected]>
| Newsgroups | gmane.ietf.mobike |
|---|---|
| Message-ID | <007201c53580$23592d20$6501a8c0@adithya> |
> On Wed, 2005-03-30 at 17:29, Mohan Parthasarathy wrote: > > When the message is sent outside the window, is it still protected > > with the IKE SA or not ? > ikev2 says you MUST NOT send a message outside the window. > My understanding is that in Pasi's MOPO protocol, the PATH_TEST message is sent outside the window and also outside the IKE SA. (Perhaps someone could clarify that). Once you know the new PATH, retransmit all the messages in the existing window in the new PATH. Then, change an explicit PATH_CHANGE message. In this method, you needed only the *PATH_TEST* message outside the window (and also outside the SA). MOBIKE messages itself was sent within the window. Jari's question was about all MOBIKE messages i guess. If all of them needs to be sent outside the window, then defining its own message-id space makes sense i guess. > > Even if the window size is greater than 1, the window may be filled > > up with previous messages. So, we are back to the same problem, right ? > > effectively, yes, unless you prereserve some chunk of the window for > last-minute mobility. > > (I suspect that we may be better off adding a distinct sequence space > for mobility requests, viewed as a slightly lower layer than the > existing phase 2 requests). > Yes, MOBIKE can define its own space and the message-id can be carried in the payload itself i guess. -mohan > > > _______________________________________________ > Mobike mailing list > [email protected] > https://www.machshav.com/mailman/listinfo.cgi/mobike