[prev in list] [next in list] [prev in thread] [next in thread] 

List:       sip-implementors
Subject:    RE: [Sip-implementors] Question about UDP/TCP switching
From:       Dasgupta Angshuman-A18655 <angshuman () motorola ! com>
Date:       2003-05-26 14:05:35
Message-ID: C73F06B054B7D6118C040008C7F3613D013C71F6 () zin09exm01 ! corp ! mot ! com
[Download RAW message or body]

The re-assembly needs to be avoided...
This will, for bad networks, increase the reliability of message delivery over UDP, \
since the entire message will be in one packet.

If your (possibly mobile) platform does not support connection-oriented transports, \
you need to keep the message size below the switch-over size. Otherwise, you might \
experience inexplicable delays/retransmissions, from the application perspective. 

Angshuman


-----Original Message-----
From: Bang Du [mailto:bangdu@ccpu.com] 
Sent: Saturday, May 24, 2003 3:56 AM
To: sip-implementors@cs.columbia.edu
Subject: [Sip-implementors] Question about UDP/TCP switching


Hi All,
In RFC 3265, if a Request message size is bigger than (MTU - 200), assuming UDP is \
used, then we need to switch UDP to TCP to prevent from message segment.

Can anyone please tell me what's the reason for this? Why is it mandatory? My concern \
is IP will be responsible for fragmenting and reassembling the datagram.

Any help will be appreciated.

Thanks,
Bang

_______________________________________________
Sip-implementors mailing list
Sip-implementors@cs.columbia.edu \
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors


[prev in list] [next in list] [prev in thread] [next in thread]