>>Usually it is simpler to implement the following strategy:
>>
>> - If the message is below 64K, send the message using UDP.
>> - If the UDP transaction times out, try TCP transport.


Simple, yes.

But bad also.

Why do people want to use UDP for large messages?
It's a great way to block things up.
And on a lossy network the problems are exacerbated.

Correct me if I'm wrong but if you have a fragmented UDP message
and one of the fragments is lost, then the whole message
has to be re-transmitted.

Also, if there are 2 clients sending large UDP messages to the
same server aren't they likely to end up blocking eachother?


_______________________________________________
Sip-implementors mailing list
[email protected]
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors

Reply via email to