From owner-freebsd-hackers Thu Oct 26 22:33:55 1995 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.6.12/8.6.6) id WAA00231 for hackers-outgoing; Thu, 26 Oct 1995 22:33:55 -0700 Received: from blob.best.net (blob.best.net [204.156.128.88]) by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id WAA00225 for ; Thu, 26 Oct 1995 22:33:51 -0700 Received: from geli.clusternet (rcarter.vip.best.com [204.156.137.2]) by blob.best.net (8.6.12/8.6.5) with ESMTP id WAA05366; Thu, 26 Oct 1995 22:33:47 -0700 Received: from localhost (localhost [127.0.0.1]) by geli.clusternet (8.6.12/8.6.9) with SMTP id WAA05178; Thu, 26 Oct 1995 22:31:24 -0700 Message-Id: <199510270531.WAA05178@geli.clusternet> X-Authentication-Warning: geli.clusternet: Host localhost didn't use HELO protocol X-Mailer: exmh version 1.6.4 10/10/95 To: "Jordan K. Hubbard" Cc: hackers@freebsd.org Subject: Re: New lmbench available (fwd) In-reply-to: Your message of "Thu, 26 Oct 1995 21:34:54 PDT." <16298.814768494@time.cdrom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 26 Oct 1995 22:31:23 -0700 From: "Russell L. Carter" Sender: owner-hackers@freebsd.org Precedence: bulk > > > Sorry folks, my ISP is downgrading from FreeBSD Pentiums to SGI > > > multiprocessor boxes (no offense Larry ;-), so guess what, I've been > > > basically hosed for the last three days. > > > > You went from FAST machines to SLOW machines? Why in earth would you do > > that? > > Actually, they're not the only ones. BEST did the same thing and > apparently the SGI is working fairly well for them.. But hey, let's > compare apples with apples here.. The SGI machines are multiprocessor > R4000 boxes costing tens of thousands of dollars, and a fully loaded > P5 system will run you $5K or so.. ahem. cough. "BEST" is my ISP... and there is more fun to come, seems we have had a (and I paraphrase here) "sophisticated and persistant" attack that has killed the SGIs in many interesting ways. All through this the lonely little P5-90 box with the NCR controller running a July kernel is just chugging along, so web service has been unaffected, as far as I can tell. > > The big problem all ISPs are having is that this stuff just doesn't > scale very well. If you could just drop another P5 into the soup and > "cluster" it transparently then we'd be looking at a whole 'nother > ballgame, but.. This kind of forces the really big ISPs into doing > exactly what Russell's ISP has done.. :-( So here is a market opportunity: how do you scale a bunch of httpd servers working in concert (maybe like timed?) so that when one goes down, the remainders elect a master and life goes on? Cheers, Russell > > Jordan >