Sad Comparison of Windows Server 2003 and Linux (RH8 and RHAS, of course)
On Friday 09 May 2003 19:59, Matthew Carpenter wrote:
Saw it on linuxtoday. The file system wasn't mounted noatime while the MS box was, it had the default block size, while the MS box was tweaked, etc etc etc All in all, the best that can be said about it is that they were honest about the funding. I'm a bit disappointed of HP though, for sponsoring this POS. Jerry, what do you have to say for yourself? :)
Matthew Carpenter wrote:
Pretty standard techniques - I've seen these commissioned third-party 'unbiased' benchmarks about twice a year with a different FUD target each time. Fundemental concept is to make competition look real bad by having a hyper-tuned product in comparison to an untuned/detuned/overloaded-with-other-junk 'other system'. Then the vendor can legally post the independant results, reference the document while keeping their own hands clean, get the press excited, have the competitor waste huge resources combating the FUD instead of developing, etc. Sad part is - the press actually laps it up. Saddest part - the poor saps who accepted the commision to do the benchmark. My only question is: why would anyone want to use a linux machine as a file server in a windows environment? The network chatter would make any samba box have a nervous breakdown.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Friday 09 May 2003 2:45 pm, Hans Forbrich wrote:
Matthew Carpenter wrote:
Pretty standard techniques - I've seen these commissioned third-party 'unbiased' benchmarks about twice a year with a different FUD target each time.
I sent a message to the author thusly: What I would like to see, however, is for folks to advocate a series of trials that are designed to "level the field" a bit: * an "out of the box" test: under careful observation that "pure defaults" are chosen for an off-the-shelf [random sampling] install, then run the tests. points deducted if any "system/kernel" level changes are needed to actually perform the test, and if system "A" needs to make a change, proponents of system "B" are allowed to make the same of functionally similar change and repeat the test, keeping the better of the two scores. * a "performance pro" test: have the top people for each respective platform pull out "all the stops" to make their system the fastest. The systems should, of course, be kept on seperate networks to avoid "interference" from one system to the other and finally, * a "real world" test: have sys admins enter a random drawing -- pick 5-10 for each system and have them do the "performance pro" test to the best of their abilities (no consulting between team members, though perhaps allow them internet access to research techniques...) The "real world" part of this isn't the software nor knowledge of esoteric settings, but rather a test of "the guy you hired last week being able to maximize the system" - -- Yet another Blog: http://osnut.homelinux.net -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.2-rc1-SuSE (GNU/Linux) Comment: http://osnut.homelinux.net/TomEmerson.asc iD8DBQE+vCWEV/YHUqq2SwsRAh5dAJ9E3TxhxTwlVd/QLTVlLED6cJeaUACgzt26 x0Z71ksy9uJB7mxc4e3Lzmk= =2H2b -----END PGP SIGNATURE-----
While I agree with the concept ... there are a few 'reality check' concerns that I've run across in the past (embedded). Tom Emerson wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Friday 09 May 2003 2:45 pm, Hans Forbrich wrote:
Matthew Carpenter wrote:
Pretty standard techniques - I've seen these commissioned third-party 'unbiased' benchmarks about twice a year with a different FUD target each time.
I sent a message to the author thusly:
What I would like to see, however, is for folks to advocate a series of trials that are designed to "level the field" a bit:
Interesting notion. Since this was commissioned (read- paid for) by Microsoft, do you think they might be interested in a level playing field?
* an "out of the box" test: under careful observation that "pure defaults" are chosen for an off-the-shelf [random sampling] install, then run the tests. points deducted if any "system/kernel" level changes are needed to actually perform the test, and if system "A" needs to make a change, proponents of system "B" are allowed to make the same of functionally similar change and repeat the test, keeping the better of the two scores.
And also read - supported by a hardware partner (no comment whether the support was willing)
* a "performance pro" test: have the top people for each respective platform pull out "all the stops" to make their system the fastest. The systems should, of course, be kept on seperate networks to avoid "interference" from one system to the other
Who decides 'who is the pro' for each team? The person paying, of course. Better to leave this alone otherwise the winner can say "and they had/did their best ..."
and finally,
* a "real world" test: have sys admins enter a random drawing -- pick 5-10 for each system and have them do the "performance pro" test to the best of their abilities (no consulting between team members, though perhaps allow them internet access to research techniques...) The "real world" part of this isn't the software nor knowledge of esoteric settings, but rather a test of "the guy you hired last week being able to maximize the system"
In the long run, it's much better to expose the fallacy.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Friday 09 May 2003 4:02 pm, Hans Forbrich wrote:
While I agree with the concept ... there are a few 'reality check' concerns that I've run across in the past (embedded).
Tom Emerson wrote:
What I would like to see, however, is for folks to advocate a series of trials that are designed to "level the field" a bit:
Interesting notion. Since this was commissioned (read- paid for) by Microsoft, do you think they might be interested in a level playing field?
Of course not :) that is why I'm saying folks like the author of the article [i.e., "the press"] should "advocate" a more realistic set of tests. [and yes, I know getting "the press" to be impartial (or even as cluefull) is another totally useless gesture...]
* a "performance pro" test: have the top people for each respective platform ...
Who decides 'who is the pro' for each team?
Actually, "the readership" of whomever is doing the reporting -- or maybe (just for laughs) it should be Bill Gates (in person) vs. Linus Torvalds (likewise). BG does (or at least did) have some "technical prowess", but I fear most of it has been directed at marketing and legal shenanigans. Likewise LT is well known for getting this whole thing off the ground, but (as I understand it) he is getting paid to do "other things", so mirco- management of the kernel and/or other test points might not be his strongest suit at the moment. [and actually, there is far more to either Linux or MS-DOS than one man can "keep up with" -- perhaps we could make them team "captains" and have THEM select the team members...] - -- Yet another Blog: http://osnut.homelinux.net -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.2-rc1-SuSE (GNU/Linux) Comment: http://osnut.homelinux.net/TomEmerson.asc iD8DBQE+vtWEV/YHUqq2SwsRAs2PAJ9R+exFv7bX1LZmj0t9qxxNa/U95ACfaMWp +DblHbVusb1nAE9eUaKl9fY= =qpAX -----END PGP SIGNATURE-----
On Sunday 11 May 2003 5:58 pm, Tom Emerson wrote:
Who decides 'who is the pro' for each team?
Actually, "the readership" of whomever is doing the reporting -- or maybe (just for laughs) it should be Bill Gates (in person) vs. Linus Torvalds (likewise). BG does (or at least did) have some "technical prowess", but I fear most of it has been directed at marketing and legal shenanigans. Likewise LT is well known for getting this whole thing off the ground, but (as I understand it) he is getting paid to do "other things", so mirco- management of the kernel and/or other test points might not be his strongest suit at the moment. [and actually, there is far more to either Linux or MS-DOS than one man can "keep up with" -- perhaps we could make them team "captains" and have THEM select the team members...]
Actually, I follow the Linux Kernel Mailing List (LKML), and I can tell you that Linus is quite active in quite a number of very technical discussions. He may not have written it all, but he sure knows enough of it to be dangerous. To describe him as being out of touch with the kernel would be, IMHO, inaccurate. -Nick
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Sunday 11 May 2003 4:19 pm, Nick LeRoy wrote:
Actually, I follow the Linux Kernel Mailing List (LKML), and I can tell you that Linus is quite active in quite a number of very technical discussions. He may not have written it all, but he sure knows enough of it to be dangerous. To describe him as being out of touch with the kernel would be, IMHO, inaccurate.
Touche' -- I guess *I'm* the one "out of touch" [not surprising, really...] but my point was that there is more to tuning a system than "just tweaking the kernel" [although I suspect I won't get much disagreement over the implied comment that Linus would do a better job of "tweaking" an entire linux system than Bill would of a similar windows system...] OTOH, who better than these two to "head up" a team of their own choosing? Food for thought -- I'll bow out now and consider my navel... - -- Yet another Blog: http://osnut.homelinux.net -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.2-rc1-SuSE (GNU/Linux) Comment: http://osnut.homelinux.net/TomEmerson.asc iD8DBQE+vyJyV/YHUqq2SwsRAkNtAKCId3//Ww/lrUWDCOPAUiAC3ZzDEACfVhdm G2UKSZwjRvTVrX8LcdRDREE= =0aDt -----END PGP SIGNATURE-----
Looks like the "independent" studies sponcered by the american tobacco industry to forstall lawsuits. In political circles we call this spindoctoring. Our industry would be well advised to avoid the corrupt methodologies of politicians. Programming and logic have more in common than politics and lying. Let us use our logic we learned when we took our first structured programming languages. We should leave lying to the experts in politics. CWSIV On Friday 09 May 2003 2:45 pm, Hans Forbrich wrote:
Matthew Carpenter wrote:
Pretty standard techniques - I've seen these commissioned third-party 'unbiased' benchmarks about twice a year with a different FUD target each time.
________________________________________________________________ The best thing to hit the internet in years - Juno SpeedBand! Surf the web up to FIVE TIMES FASTER! Only $14.95/ month - visit www.juno.com to sign up today!
Carl William Spitzer IV wrote:
Looks like the "independent" studies sponcered by the american tobacco industry to forstall lawsuits.
In political circles we call this spindoctoring.
Our industry would be well advised to avoid the corrupt methodologies of politicians. Programming and logic have more in common than politics and lying. Let us use our logic we learned when we took our first structured programming languages. We should leave lying to the experts in politics.
Too late. Over the past 10 years, our industry has become the model for the politicians. There are techniques that some huge vendors have used that would have made most politicians blush 15 years ago. It basically started with the focus-group magazines which existed entirely to support specific vendors and achieved prominance primarily through ads. By extremely well word-crafted reporting of 'independant evaluations', the vendors have been able to test and stretch the boundaries of our gullability. Nowadays we simply call it marketing.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Friday 09 May 2003 12:59, Matthew Carpenter wrote:
"Microsoft has Linux in its sights and is going to do everything it can to make the open source upstart look like an also-ran." That's nice.... M$ has Linux in it's sights. ROFL! Every cracker on the net had Windows in there sites (err. or should that be sights). The real test will be the amount of money those M$ drones convince their bosses and boards to spend in order to drive their admins nuts with patches, hacks, data loss and theft. Let's see. Last year the amount of patches for M$ servers and software needed was on the order of 1 ever 1-1/2 weeks. Also, we shant go into the amount of money, downtime, and revenue lost by companies due to stuff like CODE RED, Nimda, etc.. etc... If M$ spent as much money on building a better product as they do on marketing, advertising, and legal fees they might just make something worth while. It's just more FUD and more smoke and mirrors - and frankly it's wearing thin with more and more. Cheers, Curtis. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.2-rc1-SuSE (GNU/Linux) iD8DBQE+vFkg7WVLiDrqeksRAvBVAKCMypgOUMvNjXRiOmzeopn/IM0KHgCffwDS JAQ6ggY7cC/KUvrj9S/LmxY= =r7X1 -----END PGP SIGNATURE-----
Here's another side of the story... Is Redmond sucking wind with MS Server 2003? No apps, so lots of power Friday 02 May 2003, 09:45 MOST REPORTS on Server 2003 have been in the neutral to cautiously optimistic range, as the IT realm hunted around for reasons why someone might actually want to deploy it. According to Steven Vaughan-Nichols from Practical Technology, however, there's more than a little water in Redmond's gas tank with this latest release. source:http://www.theinquirer.net/?article=9271
On Saturday 10 May 2003 03:56, James PEARSON wrote:
Here's another side of the story...
Is Redmond sucking wind with MS Server 2003? No apps, so lots of power Friday 02 May 2003, 09:45
MOST REPORTS on Server 2003 have been in the neutral to cautiously optimistic range, as the IT realm hunted around for reasons why someone might actually want to deploy it. According to Steven Vaughan-Nichols from Practical Technology, however, there's more than a little water in Redmond's gas tank with this latest release.
"The price a corporation's paid for that is going to vary depending on their setup, of course, but the "Standard" edition costs $7000 per CPU. And guess what? Your $7000 per CPU investment is broken if you upgrade to Microsoft's latest and greatest. So if you deployed BizTalk on five systems you just dropped $35,000 for less than a year of functionality (BizTalk Standard was released in mid-June of last year)." This is the kind of mindless crap that M$ is doing to kill their own business. I mean what morons are going to fork out this kinda cash, especially in this economic climate, for these morons that are trying to sell this crap! Seems someone at M$ is living in La La Land. I mean their pumping out backwardsly incompatibile software, with no peripheral or ancillary program support at fairly heavy prices evey 12 to 18 months. I don't think these guys get it. Seriously, I guess because Bill and Steve are some of the richest guys around they think that everyone else can just kick out a few $100k every year on IT spending just because M$ rolls out a new product. Consider the expense to a small to mid-size business. What response do you think most IT division heads will get from their bosses if they propose spending more of their budget upgrading to WS2k3 and then justifying upgrading the newer software in order to maintian compaitiblity and functionality? Not to mention that some of your software/programs may not even have a version that's compatible with WS2k3! Is it me or does M$ seem to be getting dumber by the minute? Hmmmm! Since they can't blatantly bully business and partners into doing things they way they want they now have to actually have a business strategy and product line. They don't seem to be doing that very well IMHO. Cheers, Curtis.
Read their "Disclaimer of Warranties; Limitation of liability:" In a human non-legal language it says that if these M$ paid guys wrote fraudulent or bogus report, they're not liable for it 'cause M$ paid them. Total baloney and rubbish! Alex On Friday 09 May 2003 10:59 am, Matthew Carpenter wrote:
participants (9)
-
Alex Daniloff
-
Anders Johansson
-
Carl William Spitzer IV
-
Curtis Rey
-
Hans Forbrich
-
James PEARSON
-
Matthew Carpenter
-
Nick LeRoy
-
Tom Emerson