>Here's a preliminary agenda for our meeting in Montreal.
>Additions/changes/comments appreciated.
>4. Potential New WG Items
> * DTLS over DCCP, draft-phelan-dccp-dtls-00 (Tom, 10 min)
> * RTP over DCCP, draft-perkins-dccp-rtp (Colin, 15min)
> * DCCP Mobility, draft-kohler-dccp-mobility (?, 10 minutes)
I have another potential new WG item that I would like to talk about:
TCP Friendly Rate Control (TFRC): Protocol Specification,
draft-floyd-rfc3448bis-00.txt.
Five or ten minutes would be fine. Sorry for not getting this to
you earlier.
I am assuming that this would best go in dccp instead of tsvwg...
- Sally
>From the draft:
Changes from RFC 3448:
* Incorporated changes in the RFC 3448 errata:
- "If the sender does not receive a feedback report for
four round trip times, it cuts its sending rate in half."
("Two" changed to "four", for consistency with the rest
of the document. Reported by Joerg Widmer).
- "If the nofeedback timer expires when the sender does not
yet have an RTT sample, and has not yet received any
feedback from the receiver, or when p == 0,..."
(Added "or when p == 0,", reported by Wim Heirman).
- In Section 5.5, changed:
for (i = 1 to n) { DF_i = 1; }
to:
for (i = 0 to n) { DF_i = 1; }
Reported by Michele R.
* Changed RFC 3448 to correspond to the larger initial windows
specified in RFC 3390. This includes the following:
- Incorporated Section 5.1 from [RFC4342], saying that
when reducing the sending rate after an idle period, don't
reduce the sending rate below the initial sending rate.
- Change for a datalimited sender:
When the sender has been datalimited, the sender doesn't
let the receive rate limit it to a sending rate less than
the initial rate.
- Small change to slow-start:
Changed so that for the first feedback packet received,
or for the first feedback packet received after an idle
period, the receive rate is not used to limit the
sending rate. This is because the receiver might not yet
have seen an entire window of data.
* Clarified how the average loss interval is calculated when
the receiver has not yet seen eight loss intervals.
* Discussed more about estimating the average segment size:
- For initializing the loss history after the first loss event,
either the receiver knows the sender's value for s, or
the receiver uses the throughput equation for X_pps and does
not need to know an estimate for s.
- Added a discussion about estimating the average segment size
s in Section 4.1 on "Measuring the Segment Size".
- Changed "packet size" to "segment size".
