|
The branch ID can only be usefully
inspected by the same proxy that generated it and thus its format is
implementation dependant and not specified in the RFC. The example you quoted is
just an example, which consists of two parts - the hash of request URI and
possibly some other fields (used for loop detection) and the identifier of
the branch taken (to match responses received from multiple branches after
forking).
BTW, this
type of questions should only be posted to sip-implementors list, not to SIP
WG.
---
Igor
Slepchin
the branch field in the via header has two parts.
according to the rfc2543, the first part is for
loop check:
The "branch" parameter MUST depend on the
incoming request-URI to distinguish looped requests from
re-quests
whose request-URI is changed and
which then reach a server visited earlier.
since request-URI is only information to check loop by
branch, so the first branch part only encode request-URI ?
the second part seems for seeking the corresponding
request. so Which information should be encoded inside?
a branch in some document such as
7a83e5750418bce23d5106b4c06cc632.1. Which part is for loop-check?
which part is for matching a previous request? why the
second part usually only has small information?
thanks,
roland
|
