[e2e] Is a non-TCP solution dead?

[Mark Allman]

> At the risk of stumbling over an implied irony:) On the surface it
> it would seem that if one were to mandate no delacks, you wouldn't
> need ABC in the first place...

As discussed in RFC2465, ABC still has security benefit in that
case.  As Savage, et.al. showed in their 1999 CCR paper a receiver
can split ACKs and induce the sender into increasing cwnd faster
than it should.  So, if the receiver gets one packet and sends 4
ACKs (each for one-fourth of the payload of the incoming packet) a
packet-counting sender will increase cwnd by 4*MSS rather than 1*MSS
-- which is the intent of CC.

allman


--
Mark Allman -- BBN/NASA GRC -- http://roland.grc.nasa.gov/~mallman/