Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.ecn.uoknor.edu!news.wildstar.net!news.sdsmt.edu!news.mid.net!newsfeeder.gi.net!newsfeed.internetmci.com!news.mathworks.com!udel-eecis!wuccrc!not-for-mail
From: chuck@ccrc.wustl.edu (Chuck Cranor)
Newsgroups: comp.os.linux.networking,comp.unix.bsd.netbsd.misc,comp.unix.bsd.freebsd.misc
Subject: Re: TCP latency
Followup-To: comp.os.linux.networking,comp.unix.bsd.netbsd.misc,comp.unix.bsd.freebsd.misc
Date: 8 Jul 1996 14:04:03 -0500
Organization: Washington University, St. Louis MO.
Lines: 57
Message-ID: <4rrm33$oor@dworkin.wustl.edu>
References: <4paedl$4bm@engnews2.Eng.Sun.COM> <4rfkje$am5@linux.cs.Helsinki.FI> <31DC8EBA.41C67EA6@dyson.iquest.net> <4rqcsk$ff8@fido.asd.sgi.com>
NNTP-Posting-Host: dworkin.wustl.edu
Xref: euryale.cc.adfa.oz.au comp.os.linux.networking:44384 comp.unix.bsd.netbsd.misc:3967 comp.unix.bsd.freebsd.misc:23061
In article <4rqcsk$ff8@fido.asd.sgi.com>,
Larry McVoy <lm@slovax.engr.sgi.com> wrote:
>Finally getting to latencies et al. I think that everyone should print
>out the two long messages from Linus in this thread. In over ten years
>of working in the OS world, i have never seen a better treatment of
>the issues.
Hi Larry, hmm, I bet you dislike reading academic papers as much as I
do :-). At your suggestion I re-read <4rfkje$am5@linux.cs.Helsinki.FI>
(can't seem to find the second message you are referring to) and I came
to the same conclusion that I did the first time I read it: the analysis
presented is too simplistic because it doesn't take into account the
issue of scaling and performance under load.
Writing stuff like "Think Quality (latency) vs Quantity (throughput)"
implies that we are looking at a problem that only has two dimensions,
and that just isn't true. Isn't this one of the motivations behind
lmbench 2.0?
If you are interested in reading some quality papers in a related area
that address scaling, you might consider reading some of Jon Turner's
work on ATM Switching (http://www.arl.wustl.edu/~jst). Over the years
I've noticed that if you present something to Jon and don't address
the scaling issue he will be sure to ask you about it in the Q/A session.
On a somewhat related note, I also have a comment about the message
(or .signature?) that started this thread. Posting performance
numbers without analysis or indicating the context in which they
were gathered is completely useless. And appending "beat that" to
your numbers only indicates that you subscribe to the macho "my dick
is bigger than yours", "my dad can beat up your dad", and "your OS sucks"
philosophy which is all too prevalent on USENET and in the OS area. I
instantly file such messages in the bit-bucket along with the standard
flames ("Linux vs FreeBSD", "OpenBSD vs NetBSD", "Linux vs Lignux", etc.).
I also think posting such statements generally lower your credibility
because people remember what you've posted in the past and are less
likely to take your future postings seriously.
If anything, we need more analysis of the numbers. Just collecting a
matrix of benchmark results is not that useful [not to pick on Larry,
but look at http://reality.sgi.com/employees/lm/lmbench/lmbench-summary].
I want to know *why* the numbers are different and what can I learn from
the results. That info is useful because I can apply it to my future
hardware or software work, and thus avoid others' past mistakes, and
take advantage of others' successes.
chuck
--
>>Chuck Cranor, Graduate Student, Computer and Communications Research Center<<
>>Washington University, St. Louis MO http://www.ccrc.wustl.edu/pub/chuck <<
... help! my wife has accepted a job with at&t research in new jersey and
now i've got to find a job in new jersey too ...