Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!nntp.coast.net!news-res.gsl.net!news.gsl.net!news.mathworks.com!newsfeed.internetmci.com!zdc!zdc-e!szdc-e!news From: "John S. Dyson" <toor@dyson.iquest.net> Newsgroups: comp.os.linux.networking,comp.unix.bsd.netbsd.misc,comp.unix.bsd.freebsd.misc Subject: Re: TCP latency Date: Mon, 15 Jul 1996 08:02:11 -0500 Organization: John S. Dyson's home machine Lines: 59 Message-ID: <31EA4153.167EB0E7@dyson.iquest.net> References: <4paedl$4bm@engnews2.Eng.Sun.COM> <31E106AF.41C67EA6@dyson.iquest.net> <4rvmtf$ven@linux.cs.Helsinki.FI> <31E3D9E2.41C67EA6@dyson.iquest.net> <4s5bl2$qpg@linux.cs.Helsinki.FI> <31E664EB.167EB0E7@inuxs.att.com> <4s67sk$oa9@fido.asd.sgi.com> <31E6B8AB.3E6C@indy.celebration.net> <x791cmo9cs.fsf@oberon.di.fc.ul.pt> NNTP-Posting-Host: dyson.iquest.net Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable X-Mailer: Mozilla 3.0b5aGold (X11; I; FreeBSD 2.2-CURRENT i386) Xref: euryale.cc.adfa.oz.au comp.os.linux.networking:45261 comp.unix.bsd.netbsd.misc:4054 comp.unix.bsd.freebsd.misc:23604 Pedro Roque Marques wrote: > > John, > = > what are the precise mechanisms and design decisions in BSD networking = that > make it's TCP scalable ? > = Good question. I would like to rather focus the argument on how the general conclusion that the Linux TCP latency is faster than FreeBSD by seeing that the latency under no load is faster. I find that to be interesting considering that no such data is presented. It was also curious that a data point of three (3) connections was presented as a loading benchmark. At least that is the basis of my position that excessive claims are being made based upon benchmarks results that are presented publically. I would like to see some details actually show that the claims are substantiated =2E > > Can you argue that BSD TCP is inherently more scalable that Linux's ? O= r even > that is prepared for large servers ? > I believe that the issue has been that benchmarks that show performance under no load do not indicate performance under heavy load. A good example of this is the old (1.2.X) context switching performance. Works really well until you run 20 or so processes. Same idea. > = > (please do focus your answer on TCP) > = Well, let's say that you have internal data structures that map the IP/Port addresses and protocol to a connection. Those data structures can be hashed or a single linked list or some combination inbetween. There are other things that affect scalability (my VM work has shown several areas that are applicable to most areas of the system.) We can discuss that later. Suffice it to say, that there are significant scalability issues, and it is not clear that a claim of "superior TCP latency" has been substantiated. Additionally, the only data presented so far has show that there is a difference in performance between the drivers or driver interfaces. Otherwise, the TCP performance differences between FreeBSD/Linux under no load ar shown to be nil. However, I am not making a claim that FreeBSD is faster/slower in the networking area than Linux. I am making a claim that the benchmark results do not correspond to the claim being made... Simple as that. It is mostly an issue of truth in advertising and integrity. John