This also sounds good to me if VoIP applications use CCID2 instead of CCID 3.
-Arjuna
On 11/6/06, Sally Floyd <[EMAIL PROTECTED]> wrote:
Another possible variant for CCID 2, in addition to allowing
slowly-responding AIMD
parameters, would be to specify a variant of CCID 2 for small-packet
flows.
This would be a little more more work than the slowly-responding AIMD
option, but
wouldn't be so bad to specify or implement, and would be a parallel to
TFRC-SP.
In particular, it would consist of the following:
* An enforced 10-ms interval between packet transmissions.
* Window increases like those for CCID 2 flows with 1460-byte segments.
* Taking into account packet header size when sending packets.
(E.g., the congestion window would cover not just segment size, but
would cover segment + header size.)
This could be used with the regular AIMD parameters for CCID 2, or with
the slowly-responding AIMD parameters. It would be orthogonal.
Worth doing, or not?
- Sally
http://www.icir.org/floyd/
--
Electronics Research Group
University of Aberdeen
Aberdeen AB24 3UE
Web: www.erg.abdn.ac.uk/users/arjuna
Phone : +44-1224-272780
Fax : +44-1224-272497
