Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 12 Jul 2000 10:50:27 -0700
From:      "Kevin Oberman" <oberman@es.net>
To:        James Wyatt <jwyatt@rwsystems.net>
Cc:        Luigi Rizzo <luigi@info.iet.unipi.it>, Joao Pagaime <jpsp@rccn.net>, freebsd-questions@FreeBSD.ORG, freebsd-isp@FreeBSD.ORG
Subject:   Re: Enable SACK 
Message-ID:  <200007121750.e6CHoRn01417@ptavv.es.net>
In-Reply-To: Your message of "Wed, 12 Jul 2000 12:01:11 CDT." <Pine.BSF.4.10.10007121158220.72994-100000@bsdie.rwsystems.net> 

next in thread | previous in thread | raw e-mail | index | archive | help
> Date: Wed, 12 Jul 2000 12:01:11 -0500 (CDT)
> From: James Wyatt <jwyatt@rwsystems.net>
> Sender: owner-freebsd-questions@FreeBSD.ORG
> 
> Ok, I'll bite. It won't be the first time I'va asked a question and
> learned something from this group of folks... Feel free to answer in
> private email or post to the list(s).
> 
> What is SACK and what is good/bad about it? I don't intend to start a
> flame war, but really would like to know. I think I can go read RFC1323 
> and get the gist of it this afternoon... TIA - Jy@

SACK is a TCP enhancement to prevent a common problems of connections
oscillating in speed between line rate and near zero when several
packets are lost in a single transmission window. It is NOT part of
RFC1323 which covers a number of other enhancements in TCP to enhance
performance over very high speed links. It is covered in RFC2018, "TCP
Selective Acknowledgment Options", and is a standards track RFC.

I'm quite surprised to hear that FreeBSD does not support it as this
is a serious problem on large latency links at OC-3 and above. It's a
really big issue at higher speeds. When we did "fat pipe" testing over
an OC-12 from Berkeley, CA to Chicago, IL, we had to enable SACK along
with the 1323 stuff on the Suns that were used in the test to get
over 300 Mbps on a single TCP stream. I hate to think that Sun does TCP
better than FreeBSD.  (Compaq, IBM, and Sun all do SACK.)

R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman@es.net			Phone: +1 510 486-8634


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-isp" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200007121750.e6CHoRn01417>