Re: [tcpm] rfc2581bis-01 and RFC 2861

Thanks a lot for your comments. My comments inline.

> The draft (and RFC 2581) *allows* both and the draft *recommends* byte
> counting during CA.

True. I actually meant that but you put it the right way.

> > Now consider CBR over TCP. The window will increase quickly for the
> > first case assuming packets are smaller than MSS (100 byte packet and
> > 1448 byte MSS). It will increase slowly for the second case. For a MSS
> > of 200 bytes and packet size of 100 bytes, the cwnd increase rate is
> > 1/2 of the first case.
>
> Yes.  One could argue that the faster increase (the traditional
> approach) is inappropriate, but it is a pretty small effect (and, it can
> cut both ways---sometimes being beneficial and sometimes not).

Can you elaborate how it may not be beneficial?

> RFC 2581 bases the reduction of cwnd on FlightSize (amount of
> outstanding data) not cwnd.  So, the CBR over TCP will be reduced by
> half.  I.e., if you have 10 packets outstanding, a cwnd of 100000
> packets and take a loss your new cwnd should be 5 packets.  The notion
> of FlightSize was introduced into the spec. to accommodate cases whereby
> the cwnd does not accurately reflect the load being placed on the path.
> (As has been discussed on this list before, this is not an ideal fix
> because it causes issues in another case.)

You are right. However, as Christian Huitema pointed out, there are
fairness issues here.

Somewhere, I read about sharing information between different TCP
connections. That might be useful in this scenario to maintain fairness.

So an application like Skype when forced to send voice over a TCP
implementation (assuming UDP is blocked) which:

1) does ACK counting during CA
May increase the packet rate to increase the inflight packets to mitigate
the cwnd=ssthresh=(flightsize/2)

2) byte counting during CA
May increase the packet rate and voice packet size to MSS to mitigate the
cwnd decrease.

Regards,
Salman

_______________________________________________
tcpm mailing list
tcpm@ietf.org
https://www1.ietf.org/mailman/listinfo/tcpm