Re: [Tsvwg] draft-ietf-tsvwg-rfc2960-bis-00.txt: SACK & rcv consumption

Jon Grimm wrote:
> 
> >From Section 6.2
> 
> ....
> 
>         A SCTP receiver MUST NOT generate more than one SACK for every incoming
> packet, other than to update the offered window as the receiving
> application consumes new data.
> 
> ....
> 
> The first part of this statement makes sense to me: worst case 1 SACK is
> sent for every 1 DATA chunk.
> 
> The second part makes me wonder, as I do not see other references to
> SACK generation purely due to the receive application consuming data and
> the consequent change to rwnd (as opposed to DATA chunk ack.)  Does this
> happen (and what are the rules) or is there another situation that this
> second clause refers to?
> 

Yes, I think in TCP it does happen. It is a more rare event (I think)
and not something that happens often...

Consider the situation where the receiver side is clamped at
rwnd=0 since its application stopped reading..

The sender side goes into a loop probing occasionally

Cum-Ack=100
----DATA(TSN=107)------>
<----SACK(100 a_rwnd=0)-----
<time-out>

etc..

Now if the peer does a

<---------SACK(100 a_rwnd=32k)----

The receiver hurry's along the sender and does not
have to await a T-O for the sender to resend its queued
data.

I think this is the main situation that the "gratuitous sack" is
designed for..

R

> Thanks in advance for setting me straight,
> Jon Grimm
> 
> _______________________________________________
> tsvwg mailing list
> tsvwg@ietf.org
> http://www1.ietf.org/mailman/listinfo/tsvwg

-- 
Randall R. Stewart
randall@stewart.chicago.il.us 815-342-5222 (cell phone)

_______________________________________________
tsvwg mailing list
tsvwg@ietf.org
http://www1.ietf.org/mailman/listinfo/tsvwg