Re: [tsvwg] WGLC Announcement for draft-ietf-tsvwg-sctp-sack-immediately-02: 19th March 2013

[Randy Stewart <[e-mail]>]

Nishida-san:

Sorry for the later reply.. just catching up on my email. ;-)

in-line


On Apr 2, 2013, at 2:09 PM, Yoshifumi Nishida wrote:

> Hi Michael,
> 
> On Tue, Apr 2, 2013 at 12:27 AM, Michael Tuexen
> <[e-mail]> wrote:
>> On Apr 2, 2013, at 8:29 AM, Yoshifumi Nishida wrote:
> 
>>> 2: Because this rule will be checked on every packet transmission, I
>>> would like to know if the draft suggests to set I-bit everytime we
>>> have a chance or there might be a case where we shouldn't.
>> The sender could do it to switch off delayed SACK. However, this
>> might not be useful. Therefore, as stated, the sender should set it,
>> when he is waiting for an SACK.
> 
> OK. I just wanted to mention that in ideal equilibrium state, we might
> need to set I-bit every packet since you'll get an ACK for 1 MSS,
> which opens 1 MSS in cwnd.
> But, as far as I can think this won't be a problem as the receiver has
> a freedom to ignore it if it's necessary. (e.g. too much overhead)


I am a little confused here on the above. If I get an ACK for
1 MSS, I then increment my cwnd (say by 1 MSS in Slow start) and 
also decrease my flight size by the amount acked, 1 MSS, which then
allows me to send 2 MSS. 

In CA if I can only send 1 MSS, then I am a situation where 1 RTT
is being sent per RTT. Which means my PBA will have increased one whole
cwnd and I again should be able to advance my cwnd by 1 whole segment… so
again 2 segments are ready to send… assuming I have them to send.

Where I don't have 2 full segments I could see the i-bit but I guess I
am missing the scenario in my mind where this would happen.. could be just
me this AM ;-)

R

> 
> BTW, we might want to set I-bit when we find there's no more packet to
> send in buffer after sending a packet, even though
> SCTP_SACK_IMMEDIATELY is not used.
> Do you have any comment on this?
> 
> Thanks,
> --
> Yoshifumi
> 

-----
Randall Stewart
[e-mail]