well, there is some truth here...
if you do transaction based detection, it will work only in stateful mode.
But as the guy suspected, the TM design helps a lot with the search as
you need to look only on the hash table your requests matches.
I will have to go through the doc you mentions in order to get more ideas.
regards,
bogdan
Juha Heinanen wrote:
i asked sip list, why the via the hash is needed for loop avoidance. i
got only one response that made any sense:
i can think of 3 common reasons you might want to place a hash in the Via as
opposed to searching local transactions, which may or may not apply
depending on the implementation:
(1) speed - if you have a large number of active transactions, then it's
almost certainly a lot faster to construct and compare a hash in X number of
via headers than search all the transactions received on that particular
transport. any multi threaded design would require at minimum a read
lock on a list of transactions - doesn't sound to efficient to me!
do you think that this is a valid concern for openser, i.e., does it
take too long to find out if a pending transaction with a certain
call-id/r-uri already exists?
-- juha
_______________________________________________
Devel mailing list
[email protected]
http://openser.org/cgi-bin/mailman/listinfo/devel
_______________________________________________
Devel mailing list
[email protected]
http://openser.org/cgi-bin/mailman/listinfo/devel
