*BSD News Article 73312


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.eng.convex.com!newshost.convex.com!newsgate.duke.edu!news.mathworks.com!newsxfer.itd.umich.edu!gumby!newspump.wustl.edu!oldfart.ecl.wustl.edu!siesta.cs.wustl.edu!not-for-mail
From: chuck@siesta.cs.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: 10 Jul 1996 11:39:10 -0500
Organization: Washington University, St. Louis MO
Lines: 51
Message-ID: <4s0mbe$7el@siesta.cs.wustl.edu>
References: <4paedl$4bm@engnews2.Eng.Sun.COM> <4rqcsk$ff8@fido.asd.sgi.com> <4rrm33$oor@dworkin.wustl.edu> <4ru0p1$7e5@fido.asd.sgi.com>
NNTP-Posting-Host: siesta.cs.wustl.edu
Xref: euryale.cc.adfa.oz.au comp.os.linux.networking:44672 comp.unix.bsd.netbsd.misc:3984 comp.unix.bsd.freebsd.misc:23217

Hi Larry-


   I certainly agree that your approach of fixing the single threaded case
before moving on to the scaling case is reasonable.    


   As for quoting Linus' posting out of context: giving it a second look
I feel that there was no real need to quote it at all since we all seem 
to be in agreement that the posting in question doesn't address scaling 
issues.    At any rate, my objection wasn't with his posting, but with
my interpretation of your statement that his posting should be printed
out, read multiple times, and enshrined as The Word on the topic at hand.


   Don't look for me to defend ATM.   But, in regards to your question
about Jon Turner (who does research in ATM switching): yes, he is very
bright.   One of things that impresses me most about him is the ease
at which he can shift from working on a hardware design problem to working
a theoretical computer science problem.   It is almost scary.   You'll
have no trouble finding people to vouch for Jon's brightness.


   When I said that a matrix of benchmark numbers wasn't that useful,
I was thinking of it in the context of these newsgroups and my own personal
context.  If the point behind the benchmarking is to deliver "we suck"
or "we're great" lines to ones Corporate Masters, then there is no need
to discuss it here (or publish it on the Web?).   On the other hand, I 
think there are a large number of (most silent) readers lurking around
in hopes of Learning Something about performance tuning and benchmarking.
That is one of the reasons I find John Dyson's postings useful: although
he doesn't really talk much about the tools and techniques he uses to 
diagnoses performance problems, he at least makes an attempt to educate 
us on what the problems were and how he has attempted to fix them.   

   In other word, my goal is to improve my education in this area, and 
a table of numbers by itself isn't that useful to me.   YMMV.


   BTW, based on some of your experiences with the corporation mentality
(i.e. Sun) I would not expect them to care whether or not I like or 
enjoy using some "slow junker like solaris."    In fact, at this point, 
I would expect them to have me using either the company's own OS, or 
a Microsoft OS product (NT) if the company doesn't produce an OS.

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 ...