Jamshid, all
I'm not following this track but just a small comments on NewReno, see
below
Jamshid Mahdavi wrote:
> I've finally gotten around to reading Error, and I have a few
> additional comments on it:
>
>
> Section 2.2:
>
> > NewReno [RFC2582] apparently does help a sender
> > better handle partial ACKs and multiple losses in a single
> > window, but at this point is not recommended due to its
> > experimental nature. Instead, SACK (Selective Acknowledgements)
> > is the preferred mechanism.
>
> Both NewReno and SACK are becoming popular. I'd say that the bad
> thing about NewReno is that it takes N RTTs to recover if there are N
> losses. The good thing about it is that it requires only one side of
> the connection to implement the protocol. SACK appears to be a much
> better solution. Unless there is objection, I'd say recommend SACK
> generally and NewReno (in addition, not instead) for situations where
> SACK won't work (this would normally be because the person reading
> this would only be able to install a solution on the transmitting end
> of the connection). Here is new text:
>
NewReno Takes *at most* N RTTs to recover, because since you're
sending two packets for each incoming ACK it may be the case that
you retransmit two lost packet for the same incoming ACK.
I would also say that SACK is *generally* a much better solution. I did
some simulations two years ago and I found some scenarios in which
NewReno outperforms SACK. The reason is simple: SACK is very
aggressive and it may be the case that for heavy congestions, retransmitted
packets can be lost leading to a timeout. While with NewReno which is less
aggressive the probability of loosing a retransmitted packets may be
in some scenarios less than with SACK. Unfortunately I don't have
anymore the simulation results. But I think this makes sens and it
should also be pointed out in some papers.
Kind regards.
-- ____________________________________________________________ Omar ElloumiAlcatel Research - Network Architecture - Traffic Technology
Francis Wellesplein 1, 2018 Antwerp, Belgium
Phone: + 32 3 240 78 33 Fax: + 32 3 240 99 32 http://www.alcatel.com/crc/ ____________________________________________________________
Opinions expressed are those of the sender and do not reflect Alcatel policy or agreement
This archive was generated by hypermail 2b29 : Mon Jan 28 2002 - 09:12:24 EST