> From: Linda Dunbar <[email protected]>
> I was looking for this detail in the
> "draft-ietf-lisp-introduction-03".
Yes, I already had made a note to myself to cover this topic in the next
revision... :-) The reason I hadn't should be obvious: I think the
encapsulation is an ugly kludge, and so I prefer not to think about it!
(And I also to cover your question about the proper response to a Negative
reply...)
> does the response from MapServer have a cache timer associated? Or the
> requester can keep the response forever?
Well, _normally_ the response an ITR sees (the Map-Reply containing the
actual mapping) will come from an ETR, not from the Map-Server. (The
Map-Server just knows who is authoritative for various chunks of the EID
namespace, not what the actual mappings are. So they pass Map-Requests on to
an appropriate ETR, which replies with the actual mapping.)
I assume it's these responses you're talking about - or are you talking about
Proxy Map-Replies, sent by a Map-Server which is acting as a proxy for an ETR?
If the former, each mapping (i.e. {EID->RLOC(s)}) in a Map-Reply packet has
its own TTL (see Section 6.1.4, RFC-6830), and if this is set, the ITR should
discard/refresh the mapping after that interval. (There's also a discrete
reserved value for 'keep as long as you'd like'.)
If the latter, I _assume_ MS's set a similar timeout, but that's an assumption
on my part, I'm not sure any document specifically discusses that point.
> that there is a typo "inferfaces" in Section 13.1.
Thanks! Got it.
Noel
_______________________________________________
lisp mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lisp
