Date: Mon, 10 May 1999 14:58:39 -0700 From: "Carlos C. Tapang" <ctapang@easystreet.com> To: <chat@freebsd.org> Subject: Linux news from Client Server News Message-ID: <002801be9b30$44b2b310$0d787880@apex.tapang>
next in thread | raw e-mail | index | archive | help
The first item is of interest to FreeBSD. Somebody should suggest to Mindcraft to include FreeBSD in the "Death Match". The last item is about Ken Thompson thrashing Linux. --Carlos Linux Watch: Mindcraft Proposes Linux-NT Death Match; Microsoft Accepts the Challenge Mindcraft, in an attempt to salvage its reputation after publishing a Microsoft-funded test in which NT trashed Linux, has invited Linus Torvalds, Red Hat or anyone they choose to try to prove that Linux can outperform NT Server in what it's calling an "Open Benchmark" shootout. Microsoft has already said it will risk the challenge. The event, if it comes off, could be way more fun than anything this side of the antitrust trial. It's put up or shut up time, boys. In the controversial tests paid for by Microsoft Mindcraft declared NT to be 3.7x faster than Linux as a web server and to have 2.5x the performance as a file server. The fact that Microsoft funded the tests and that the admission was buried on page 16 of the 20-page final report (CSN No 296) set the Linux community off. The implication, and in some cases outright accusation, was that Mindcraft rigged the tests for its patron and played pawn to the mighty Microsoft marketing machine. "Mindcraft's honesty and name have been impugned," the testing house said, citing interviews with Torvalds in publications ranging from ABCnews.com to Linux Today and Salon. In particular, the Linux community was outraged by statements in Mindcraft's report that it had attempted to find the best ways to tune Linux using the standard forms of Linux support - web sites, newsgroup pleas for help and the like. It said the response was negligible. In what Mindcraft now calls an "Open Benchmark Invitation," it says that it's willing to have Torvalds choose anyone he wants to help it tune Linux, the Samba middleware and Apache web server used in its benchmarks. Red Hat, likewise, was invited to send anyone it wants to serve as Linux experts. The Linux experts can load the software and run the tests themselves if they wish. Microsoft, it said, is welcome to send its NT experts to do the same. It also carefully stated in bold text in its invitation that Mindcraft will bear all of its own expenses for the new tests, and that it will run them "at any mutually agreeable test site." Microsoft seems pretty confident. It's volunteered to turn over its own test labs for the shoot-out, the same facility where Mindcraft made its original and now-notorious findings. That location was a bit of an eye opener unknown before. The lab offer pretty much answers the question of whether Redmond plans to have its hotshot tuning experts on hand. The only direct response from the Linux community so far has come from developer Jeremy Allison, said to be the brains behind Samba, who refused to participate unless Mindcraft tests using NT clients rather than the Win9x clients it's been using so far. In an e-mail to Mindcraft he said that Samba works better with NT clients, NT better with Win9x. Mindcraft also said that it's re-run the tests on its own using Linux tuning tips that Torvalds and open source people provided. It refuses to divulge the new results until after the new tests are run, and added that the shootout tests will be run on the same hardware that it used for its own re-test. Mindcraft has set strict ground rules for the tests, which are published on its web site, but has given Linux three chance to prove its case. The Linux world gets to run a set of benchmarks using the tunes, patches and bug fixes that Mindcraft had at its disposal when it ran the second set of tests. Then it can add whatever patches and bug fixes it would like from either the Red Hat or the open source kernel.org web sites as long as they were available on April 20 when Mindcraft started its tests, but it still has to be the Linux 2.2.6 that Mindcraft tested. Finally, the Linux experts can run a set of optional tests using any Linux kernel, Apache version, Samba version, patches and bug fixes posted on the web at the start of the Open Benchmark test. There's no mention of whether Microsoft is allowed or even wants to test Windows 2000 beta 3. Linux Watch: Linuxcare Gets El Primo Backer Kleiner Perkins, the famed VC, has made its Linux bet, putting its money into Linuxcare, the San Francisco-based Linux support house. Unfortunately, nobody's saying how big the bet is but they've imported ex-IBMer Fernand Sarrat to be CEO. Sarrat has lately been at Cylink, the security and cryptography firm as president and CEO. Before that he was at IBM for 23 years, lastly at general manager, network centric computing marketing and services, which was heavy on Internet-delivered solutions. Former Linuxcare CEO Arthur Tyde has given way to Sarrat and will stay on as executive VP, focusing on operations and external relations. Kleiner along with Sand Hill Group, who's coughing up fewer bucks, are providing Linuxcare's first round. Kleiner is putting its general partner Ted Schlein on the Linuxcare board. Linux Watch: Unix Co-Creator Trashes Linux Ken Thompson, the co-creator of Unix, thinks Linux will be a short-term phenomenon. He calls the operating system "unreliable" and sees its surge in popularity as more an anti-Microsoft backlash than anything else. Thompson, still a researcher at Lucent's Bell Labs where Unix was born, lambasts Linux in the current issue of the IEEE's online magazine, Computer. "I view Linux as something that's not Microsoft - a backlash against Microsoft, no more and no less. I don't think it will be very successful in the long run," he told Computer. Linux creator Linus Torvalds called Thompson "extremely misguided." Thompson also questions how effective the open source concept is."I've looked at the source and there are pieces that are good and pieces that are not. A whole bunch of random people have contributed to this source, and the quality varies drastically," he said. "My experience and some of my friends' experience is that Linux is quite unreliable. Microsoft is really unreliable but Linux is worse. In a non-PC environment, it just won't hold up. If you're using it on a single box, that's one thing. But if you want to use Linux in firewalls, gateways, embedded systems and so on, it has a long way to go." Carlos C. Tapang http://www.genericwindows.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?002801be9b30$44b2b310$0d787880>