Slashdot Mirror
NT vs. Linux: Again
Jeff Molloy writes "The results are here
link " It's a
shame Linux didn't win, but it looks like the tests show where Linux might
have some deficiencies. Overall, it looks better than the original test, though.
← Back to Stories (view on slashdot.org)
Wouldn't more people benefit from MS management seeing to it that some QA/testing people get hired?
How about fewer new "features" and better implementation of those that are already there?
CUR ALLOC 20195.....5804M
99.9% uptime would be an MTBF of a bit over 40 days. Not exactly space-shot quality, but acceptable for e-commerce. It will have to do better to work in a telco CO, but I would bet that NT-embedded, with the unneeded bits removed, would do pretty well. It looks as if each OS has its own deficiencies, among them: NT clustering is way behind, and Linux has some key kernel bottlenecks for high volume transaction applications. Is it that surprising to find that operating systems put their pants on one leg at a time?
I wrote parts of this stuff
>Currently the VC++ from microsoft produces far superior x86 code than GCC. I'm surprized that nobody has commented on this fact yet.
Why should anyone? VC++ is basically a one-platform solution (x86) while GCC pretty much runs on everything under the sun. GCC runs on the Amiga. Does VC++? Nope. When you are dealing with a complier that runs on multiple platforms/processors the kind of optimization you are talking about can be a *REAL* headache to deal with and should be steered clear of...
It's a strategy for choosing which requests to serve and which requests to ignore for the time being- and if this is so hard to understand, I imagine the tested version of NT _is_ doing this because (a) Microsoft people are _not_ stupid, and (b) Microsoft people will always cheat given the opportunity
This is a pyrrhic cheat- you can't use it on a real web server. It has nothing to do with CPU scheduling and is purely a hack to optimize benchmarks for intranet requests
How do you know MS are doing this (cheating) - can't you just say, oh, well, our tcp/ip stack needs to be multithreaded. It seems like you are trying to mislead - introducing little hints here and there that this was all faked.
and that CPU scheduling is the only consideration in doing this, because the only algorithm that exists is serve-upon-request?
In that case, it would be interesting to see Linux against NT running a different web server. We've already seen that the bottleneck exists in Linux, even when using a different web server. Certainly, if we were to see that say, solaris kicked NT and Linux's ass, you wouldn't suggest it had something to do with sun running round cheating.
Since it was shown (dispute it as you wish) that Linux's bottleneck is it's tcp/ip stack - I don't see that your argument about algorithms has any relevance in this thread.
I'll neither agree nor disagree with that as I have no knowledge in the subject. What I will do is to offer this thought:
If this hardware really was where NT shines, what happens when linux gets tweaked to take better advantage of it? The Microsoft folks have nowhere else to go.
So, I say to you folks, take heart. Accept this setback for it is not defeat. Remember, that which does not kill us...
Give a man a match, you keep him warm for an evening.
Light him on fire, he's warm for the rest of his life
I don't care how sunny and bright and beautiful your little M$ world is. Enjoy your job fucking people over for a living and being part of one of the most greedy fucked up organizations in the world.
Why don't you just stay at winfiles.com and stay the fuck away from slashdot?
support gun control: take guns from cops
I remember a project where a person did a clean install of NT on a totally fresh harddisk, put it in a room by itself, unplugged the keyboard, mouse, network, etc. and did basically NOTHING on it. Guess what? It crashed after 53 days (or something around that). As long as you don't use an unstable kernel driver, and don't do everything as root (as some people appear to do, I must admit I do it myself...), Linux is dead stable. At least it won't crash doing nothing!
/* Steinar */
(This comment is of course GPLed.)
Microsoft will just keep inventing benchmarks that happen to
make NT look better. Nothing can be done about it, other
than observing that those benchmarks will become less
realistic every month.
The only way arround this would be for Linux (Apache and
Samba) to copy the same "unapproved benchmark veto"
clause which makes the publication of truly independant
benchmarks unlikely.
I'd rather do it with a human being(female of course). I wouldn't trust a guy touching my internal organs
Very interesting comment.
All I'm saying is that saying that Linux is free and NT costs $$$ is NOT a very good argument.
>Linux uses the forked process model to provide services to multiple users. This modem achieves stability in that if one process dies, the others continue as if nothing had happened. Both Apache and SAMBA operate in this way I believe.
>NT has chosen performance over stability.
I wouldn't put it that way... Perhaps a better way to phrase it would be: Linux has chosen a pessisimistic approach to application stability over general performance.
Linux's use of processes vs threads only has merit if you assume that the processes you are running have bugs (and will crash). It really seems to suit the open-source model to be more optimistic concerning application code and give it the benefit of the doubt along with a hefty performance boost (in the form of threading). Does anyone doubt that Apache or Zeus could use the thread model, remain stable, and thereby match NT's performance?
Oxryly
Many of us met and listened to Redhat last night at the Miami roadshow. They accepted the benchmarks as accurate. They also reminded us to be like Linus and keep a sense of humor and pespective about the whole thing.
To paraphrase Mr. Torvalds....
Microsoft is just being a good Linux user and reporting bugs. The same can be said about ZDnet.
Getting your butt kicked every once in awhile (metaphorically speaking) can be a good thing. It keeps one from becoming arrogant and complacent. It can motivate you to do better and try harder.
Just think how bad American cars would still be if the Japanese hadn't come into the market. (Not to say they are the best, but they are a hell of a lot better than they were 10 years ago).
Last night at the Miami Redhat roadshow, questions were asked about the Mindcraft benchmarks.
They said one very important thing. Mindcraft was not able to duplicate their results.
About the the ZDnet benchmarks. They happen to agree with them, but not in a negative way.
They reminded us to be like Linus who has kept a sence of humor and perspective about all this.
Accoring to the folks at redhat, Linus said....
"Mirosoft is just being a good Linux user and reporting bugs. You can say the same about ZDNet."
I beleive that ZDnet was trying to be as fair as possible. They firmly believe in the future of Linux, and have stated that publicly several times. They are doing their part in helping it to become a better OS through constructive critisism.
IMHO MS is going to lose out in the long run as long as the Linux community remains honest about it's shortcomings. No multimillion dollar spin masters to hide the warts. No FUD. Just keep getting better and better, and Linux will win.
Things have changed after these Linux-NT test ...
I would never expect anyone on Slashdot to write "
If raw speed is your monkey, then NT is the tool. "
Hehehe... Well, but we can't dispute that. Right know NT is faster.
I run W2K 24x7 as a desktop OS for development. The only crashes I've had were due to an immature NVidia TNT driver (video drivers bypass the HAL and can therefore bring the system down). I've recently installed NVidia's updated driver and the crashes have disappeared. I think M$ may yet pull it off w/ W2K -- at least they'll put out a kernel (VMM, FS, net stack, etc.) that will be stable and high performance.
Oxryly
Great enthusiasm - You sound like Steve Balmer :) The problem here is that you're wrong on several key points
;)
:P).
;) I presume the W is for widechars. ;) *side note* Microsoft is such a fascinating company to follow...it has such interesting people, like balmer, gates, allen (who has dissapeared off the face of the world recently) etc...a bunch of geeks (some more than others) becoming billionaires.
LOL
1. Win2K's interface is not improved. It sucks. NT4 was good. I get paid to admin NT4. I like NT4. Win2k is a major step backwards in usability. The ungodly number of wizards in NT5 (oops win2k) makes it impossible to do any real work. Sure you can turn them off, but the mere sight of them drives me nuts - it's like having 4000 of those fscking dancing paperclips. This is supposed to be a server os - wizards don't belong on a server os.
I have to disagree there. I think the W2K gui a slight improvement (GUI here, not tools). The new MMC is great. You can administer everything from one program (including add devices, read event logs, add users, make shares etc). What's more, you can do it to remote machines...seemlessly.
2. Win2k's performance. This sucks too. Win2k takes ages to boot. Once it's up, using office 2k takes far longer than NT4 + Office97 ever did. My box is a PII 450 128Mb of RAM, I know that's not enough for the 2k products, but the company won't splurge for an upgrade.
I think it's wonderful!!! My W2K box boots in no time, and it boots in even less time when i use the cool new hibernate (memory to disk) feature. yummy. I'm running a K6-200 with 192MB ram (did have 64MB, but it was a bit sluggish with all the services installed).
3. Stability. This is anecdotal, but I've had more lockups (5) and blue screens (1) with NT5 than I had on the same box with NT4 (3)lock and (0)BSOD - Admitedly it's still in beta
Wow, I haven't had any bluescreens except one where i installed an unsigned NT4 driver i was warned not to install..after that, everything else was perfect...been running for weeks with no problems with BSODs (it's more purple now tho
4. Ease of development. There is a special place in the most fiery pit of hell for someone who names a function RegisterServiceCtrlHandlerW() Don't tell me that Win32 makes life easier for developers. It spawns carpal tunnel is what it does Again i disagree, Windows is the most develop friendly OS...even 90% of *those* java developers use Windows. It's got brilliant IDEs, which make up for the long API names, but remember, VC++ has intellisense, so you don't have to spend too much time typing, or going round documentation trying to remember what arguments you need to pass.
I'd glady have long function names, than horrible IDEs without intellisense! Besides, RegisterServiceCtrlHandlerW makes perfect sense
as for your experience at MS...uh,
Its time us non-M$-haters come out of the wood-work on /. Don't get me wrong... I love the open-source movement and approach.
But meanwhile though I run W2K beta 3 as a development system to be productive and it does everything I need it to do, quickly and without bugs or crashes. Harrumph.
Oxryly
I do remember. Microsoft was right about that one...
Actually, Microsoft does not fear re-use of old PCs with Linux. If anything, this reduces the culture of software piracy in the developing world. This is also a tiny subculture compared with 10 million new PCs manufactured every month (more than TVs now). This is, as far as I can tell, the central story of Linux. It has been covered. And it is well known by most people. Microsoft should fear, and does fear, anything that might catch up to them in terms of the total value proposition, including performance, reliability, capabilities, available applications, etc. in markets where money can be made.
I wrote parts of this stuff
Nah, that's the wrong way to go... Notice that the 'low-end' system was still spankin' new hardware which probably ran well into four digits. The lovely thing about Linux is that you can run a web server (not slashdot, I'm sure, but a web server) on a sub-$1k box. Obsolete hardware is obsolete no more. _This_ is Linux's main source of strength.
Of course, that doesn't mean it wouldn't hurt to beef up the scalability, and that's what's planned for 2.4/3.0 anyhoo...
-grendel drago
Laws do not persuade just because they threaten. --Seneca
I may have disliked the tone of the Mindcraft test, but they did provide the documents for the entire test configuration. For personal satisfaction, when can we see the equivalent documents from this PC Week test ? I don't think PC Week is hoilding this information back for any other reason than they forgot how intrested most of us are in the details". PC Week... publish the details. Please. :)
Okay, try Interbase. I'ts much faster, and works on Linux/NT and others.
"...I think that it's safe to say that noone in the world gets more than 150 million hits per day of static content..."
and
"...Oh, if you're serving up >1800 files per second of 2k files, who are you?..."
Flying Crocodile, Inc. (www.flyingcroc.com). We have FreeBSD/Apache machines running at 27Mb/s. This is on off the shelf Pent-II boxes, ok, the have 1GB of ram and UW-SCSI drives... but single cpu boards. With the new FreeBSD 3.2-RELEASE the above mentioned box runs with a load of 3.3 and over half the cpu idle.
The numbers that you talk about are the numbers that I deal with everyday. We would never think of running NT. With over 130 servers we couldn't afford the massive staff to sit around and reboot the boxes all day and night... that and I would hate to have to wire a monitor/keyboard/mouse to each box!
The servers together do over 145M hits per day and thank god the Cisco GSR12008 is shipping next week, the three 7507's are hammered!
"...Oh, one more thing. If this is all on an intranet, you'll still need Gigabit ethernet if you're serving up the 10k+ files..."
More like 2 GE to Frontier and Teleglobe, with various other T3's get the job done. BTW: before you start crunching numbers, not all the 130 servers are cranking 27Mb/s, many are doing massive database work.
Another interesting number, at our peak in the day we route about 79,000 packets per second, figure that about 1/4 of those are http requests. Peak total for today was 419Mb/s using mrtg.
IMHO: I use to be a Linux nut, still use it for desktop work, but FreeBSD kicks ass when it comes to serving. If you think my numbers are crazy, Yahoo trucks twice the bandwidth, no wonder they use FreeBSD too.
The tests have been done by those who's bussiness is to crank out the hits. Most use neither NT nor Linux.
If you doubt this post, just do a looking-glass on 207.246.128.0/20... the connections exist.
-- I have a private email server in my basement.
And the emperor (*place-any-ms-competitor-who-is-winging-to-the-go verment-here*) was really the bad guy.
It did mention that they did try zeus, the other web server suggested by linux advocates. But the cut off point was yet again present. The source of the performance block was the tcp/ip stack, not the web server.
HAHAHA!!!! yeah.
at my UNM, just the SRC computer pod with 16 clients seems somewhat confused if you do something rash on one of them, like dare to open netscape or something. NT sucks.
Juln
I get 40 days (41+2/3) for MTBF only assuming it will then be down for a whole hour: not a valid assumption I guess. Assuming a 5 minutes downtime results in the 3-4 days mentioned earlier.
Yes yes... this test did not consider every possible variable for every conceivable hardware setup for all time for all people etc. etc...
My point is this test only set out to show that given a certain hardware setup with excellent *theoretic* performance for an interesting task (web and file serving) the OS and application setup that gives the best *actual* performance is XYZ. In this particular case XYZ happened to be NT with IIS.
There are a significant number of supplementary issues that any potential OS customer must consider in addition to the information derived from the results of this test. That should in no way detract from the importance of the results of this test.
And importantly, as members and potential contributors to the open-source movement the results of this test give us an excellent report card on the progress of Linux development. I really don't think the test of the results should be excused away for any reason.
Oxryly
Multithreading is a good design -- sometimes. Multi-tasking is a good design too -- sometimes.
When two or more tasks legititmately belong in the same address space (e.g. must manipulate the same objects in memory), sure, multithreading is the way to go.
When you're just kludging a substitute for fork() because spawning processes is too expensive on your architecture, that's not so good from a stability or security standpoint, although spawning new threads.
None of this has anything to do with lack of multithreading in the Linux IP stack, which undoubtedly would be a good thing.
Post may contain irony: discontinue use if experiencing mood swings, nausea or elevated blood pressure.
Well it was almost 3 years ago, but bsdi did this.
Tests were done on P133's with 64M ram and BSDi
walked all over NT.
http://www.BSDI.COM/press/19960827
Would be nice to see a rematch.
I just read a Web-Benchmark in c't (http://www.heise.de).
They used a SMP-Patch. And then Linux was in every test at least as fast as NT, if there is only one network-card in the computer. If they used two network cards NT was about 100% faster than Linux. NT was very, very slow when using Perl-scripts.
They showed also a test on Mac OS X. The results shows very good performance of the Mac OS X although the Mac hardware was no comparable with the hardware used for NT and Linux system (i.e. 128MB RAM in Mac and 2GB in NT/Linux-Server). But it seems, that Mac OS X has a bug which causes a "system panic" when using special CGI-Scripts.
Been working with NT (unfortunately) for five years now. SMB is integrated with the Server Service, which is in kernel-land. This explains the speed.
They claimed that if the Linux box were tweaked, Linux
/dev/null)?
would win... which it did not.
Has the full configuration and "treak list" been published at
all. AFAIK only the fact that the Linux team were specifically
forbidden from performing certain tweaks.
Also were any of the tests carried out with ordinary NT
which is a far better match to RedHat 6 than NT Enterprise...
Or maybe someone should have got them to build a
"RedHat Enterprise" everything compiled for these
high end machines. (And incapable of running on
low end machines.)
Another of the original issues was logging, if this is
being done properly then every SMB or HTTP connection
will generate a synchronous write to disk. This will
slow things down. AFAIK NT dosn't log anything to do
with file shareing by default. In the original tests IIS was
placed in a mode of buffering the logging information
and writing in chunks. (In the real world you may as well
turn off the logging all together as use this option.) Did
Apache and Samba have all logging turned off (or
directed to
Hi folks!
...), but the main findings of the Mindcraft study are true under the given test circumstances.
I think one has to accept that at the moment NT is slightly faster considering the maximum output of a web or file server.
But as German PC weekly "ct" found in their own benchmarks (issue 13/99, pp. 186), Linux is still a very good choice under real world conditions. They tested SuSE Linux 6.1 and NT 4.0 SP4 on a 4-XEON-450 Siemens machine. The main difference between their configuration and the Mindcraft one was that they just had one Ethernet card (instead of FOUR!)in the system.
They said it was not realistic (except for a few Intranets maybe) that a web server has to serve more than 100 MBit/s or even more than 10 MBit/s. Under these circumstances Linux was slightly faster with static web pages and much faster with serving CGI. However, ct didnt test MS IIS with ASP (hard to find a fair benchmark between Perl/CGI and ASP anyway).
Only when they tested the system with a second Ethernet card, simulating similar loads to the ones in the Mindcraft tests, NT was significantly better (and scaled the CPUs much better than Linux.
What they also found out is that NT was much worse with serving from the HD instead of the memory (maybe because they also used one big partition instead of smaller ones, which seems to slow down NTFS. The bottom line: Linux with Apache is a very suitable and fast system for real-world (mid-size) web serving needs, mainly if you have to deal with a lot of dynamic pages (like on Slashdot
Posted by FascDot Killed My Previous Use:
Why aren't they back-porting that multi-threaded IP bug thing to 2.2?
---
Put Hemos through English 101!
OK Kernel guys (And people who have setups who can
test this sort of thing). We now know where the problems are (OK We already did) Lets fix them and then challenge for a rematch.
Erlang Developer and podcaster
I think Linux will come out ahead in the long run. It's only a matter of time.
-- DuckWing
It does disturb me somewhat to see that Linux loses on the single proc box, but this seems to come down to the tuning. Out of the box, Linux is the faster (as other benchmarks have illustrated), but when tuned, NT is better.
I think they ought to make this an annual competition and see how they match up every year. I bet next year the results won't be so slanted in MS's favor.
---
This sig has been temporarily disconnected or is no longer in service
If you can't help program then go out and test all this new stuff and send in bug reports. Let's have Linux set the standard again. It seems like, acording to the article, it was this way once and we lost it because Microsoft has pushed the bar a little higher and we lagged behind.
IIRC Linux once said:
"Linux has a micro kernel. There are only these things in it, which are needed".
your argument would carry a lot more weight if you didn't hide behind the "AC" posting. Register @ Slashdot and people might then read what you say.
---
Posted by ChristianC:
The argument over whether NT or Linux is faster is completely irrelevant. The fact is Mac OS X Server is seven times faster than NT and three times faster than Solaris, so its bound to be faster than Linux.
Seriously, trying to help a fellow newbie--have you done the pnpdump/isapnp thing for the modem and/or soundcard? worked for me (but I had to hunt for it) with RH5.2 and 6.0. RH6.0, btw, detects isapnp.conf settings during boot--you don't need to run any init scripts anymore. Just make sure that you pick the right resource settings (and telly your mb bios that you are NOT using a PnP OS or comparing your win irq/dma stuff will only confuse you).
---
Unfortunatly, I'm wasn't referring to anything so fancy. More just being sarcastic, because the throughput difference between Apache and IIS is hardly ever going to be the deciding factor.
(In the largest NT/IIS setup I've seen, there were three actual web servers. They were 'clustered' only on the switch level. The assumption was that one of servers would be down at any given point in time. A desktop box was running software which checked if IIS was running, and if it had died, attempted to restart the service. If that failed, it rebooted the box.)
--
Business. Numbers. Money. People. Computer World.
"how many times did they have to reboot NT during the test?"
None.
But, you wouldn't expect to either. Linux's stability advantage over NT is shown over weeks and months, not hours or days.
Steven, Senior Technology Editor, Sm@rt Reseller
You're mistaken.
Steven, Senior Technology Editor, Sm@rt Reseller
Chas, you're a goddamn lamer. Shut the fuck up already.
Yeah. Whatever. Great argument.
Chas - The one, the only.
THANK GOD!!!
Chas - The one, the only.
THANK GOD!!!
Please to be pointing out the PhD theses written by any of them?
And what in hell does this have to do with coding ability? Most of the best coders I know don't even have a bachelor's degree.
(And no, I'm not talking about @ Microsoft here - I'm talking about other people in the industry)
If you ain't got the spark , you may as well forget it. You can be a PhD in CompSci and not have a clue about how to code shit-hot stuff.
Coming soon - pyrogyra
This one's fun. Go to investor.msn.com and get the historic chart data for microsoft stock. Then click in the middle of the chart, hold down the primary mouse key and move your mouse side-to-side. Within 5 seconds you get taken to blue screen city. This is 100% reproducible on my W2K beta 3 box. I am not kidding. I thought MS servers were only supposed to crash your machine if you were running netscape ;)
Of course! Having comprehensively *lost*, what else can they do?
But do not think that MS will stop and wait for Linux to catch up.
This has been a public information announcement.
I also want to see vanilla installs of the two competing
operating systems... Commpared to when they both are
tweaked.
How about testing between
a) pre installed NT box
b) pre installed Linux box
c) Retail NT (bought anonymously) installed on blank machine
by NT experts. Only basic configuration allowed.
d) as c with Linux experts
e&f) as c&d using a retail Linux distribution.
g,h,i&j) as c,d,e&f but can tweak using only publically
available information, all sources and tweaks to be fully
described in the final report.
k) the NT team can do any tweaks they like subject
to a budget
l) ditto for the Linux people.
Despite the facts and fiction that have been said about the comparison of Linux to NT, something else has surfaced: most Slashdot readers seem te be reading only Slashdot. It is so strange to see how people seem to depend only on this otherwise fantastic source of info. It is only ONE source. (But it's OPEN! harr harr.) A few days ago I read an article in a German computer magazine - sorry, THE German computer magazine - and in it they showed that the tests Mindcraft (what a stupid name BTW) were everything but real-life. And because the guys at PC Week Lab's did the same tests, their tests suffer from equal non-life-realism (does this word exist in English? Nah, don't matter.).
In the German article, Linux did pretty well (on a 4CPU / 2GB RAM / RAID-5 server). Even SMP wasn't bad all the time.
I realise that German magazines, especially the dead-tree variants, are not as accessible to most Slashdot readers. But I still get the feeling that if I were to say "BOO!" on slashdot, ten minutes later all hell would have broken loose.
Now, the moral of the story:
- The Linux-NT comparisons of both Mindcraft and PC Week Labs lack realism.
- Slashdot readers: think before you send comments.
They didn't, however mention the fact that they formatted the
fileserving partitions into 4 separate partitions to improve
WinNT's performance on the front page, did they?
They didn't mention much at all about the set up. Assuming
this "4 partition" claim is correct then it adds to evidence
that the NT operates as a "semi-cluster". Therefore the
comparison is meaningless. Unless you pull out 3 of the
network leads, then do the measuring....
Well. It has 4 ethernet cards, 4 CPUS (and NT ties one card to each CPU) and 4 partitions.
Thats basically 4 seperate computers inside one box under the control of one operating system...
Whats your definition of a cluster?
Personally, I don't really beleive its a cluster, but the fact is, who is really going to have 4 CPUS, 4 eth cards, and 4 partitions to split everything up?
> I guess you didn't read the article at all.
> Linux has a problem! The TCP stack is single
> threaded and can't compete with NT on multi-proc
> systems
Actually, I did read the article. I'm just suspicious. For all I knew at the time, the explanation regarding the TCP stack could have been false or distorted. Later things that I read seemed to show otherwise, that indeed the TCP problem was a real and legitimate problem. Fair enough.
I read an article later at the Linux Daily News (6/27/1999) analyzing the results of not only the benchmarks in question, but previous ones, including those that had Linux 'winning', focusing on differences in the testing conditions. The number of clients really made a difference; with a large number of clients, Linux's performance sank due to some bottlenecks like the TCP problem, which is partly why the Mindcraft rematch came out as it did. Also fair enough.
The trouble was that it looked to me at the time that PC Week was merely detailing Mindcraft's results, which didn't tell much at the time since I knew Mindcraft had done hatchet jobs before and had a vested interest in the outcome. They could very well have made up a technical explanation as to why the results turned out as they did. To determine whether the results were due to dishonest testing or to real problems, or even both, requires having others look at the results (which has since been done). The point I was trying to make was that Mindcraft shouldn't have been taken at their word, since they haven't been all that honest before.
that's called a geesh. don't ask why. those are the rules...
Posted by ChristianC:
The argument over whether NT or Linux is faster is completely irrelevant. The fact is Mac OS X Server is seven times faster than NT and three times faster than Solaris, so its bound to be faster than Linux.
so if the CPU is idle while the network i/o is limited does that mean that handling cpu expensive stuff like e-commerce (mod_perl,zope,php & db2/oracle/m(y)sql) will perform well because the networking stuff won't be so close to its limit?
A.
Intranet/Internet Developer & Linux Advocate
Microsoft wouldn't have started this, if the result wouldn't have been predictable.
I could write here a long list, why these simple brute performance tests are not appropriate to compare operating systems. We know now that Windows NT serves the Windows native network file system better on higly optimized machines and a
NT machine can flood your network with twice the amount of static web pages than Linux.
Microsoft defined with these tests the battle ground. Next time we should know the territory before going into the battle.
It is such a shame that MS (and customers!) put such little value on reliability. Only yesterday (25/06,) I had problems with my bank, who have gone completely to NT. Account records were down for hours! If it hadn't been for the irony of the situation, I would have been furious, as I had issues I had to resolve.
I'm dreading Y2K.
BTW, the bank was NatWest (www.natwest.co.uk.)
Similar experiences here... but if there is no BSOD, and the machine just freezes, although it should be hardware, and is generally solved by changing hardware components, why does Linux behave fine on the same systems?
Maybe NT is pushing the hardware more, maybe NT is more succeptable to being affected by certain types of hardware failures, but in my experience Linux is more stable than NT.
I've been toying with the idea that maybe linux kernel mode processes are statistically a "smaller target" than NT kernel mode processes. That failures in Linux are far more likely to occur in userspace... and leave the system perfectly functional. This is going way beyond the area of my expertise, it sounds positively nutty, but it's a working theory.
System lockups are almost certianly hardware related, but the same hardware failures just don't seem to happen on Linux. I can get a rock-solid linux box from almost any cruddy hardware. I just can't do that with NT.
Here's an experiment... Let's get a rock-solid computer, and generate a kernel-mode process which will randomly shuffle a bit somewhere on the system, simulating a botched memory cell or a hardware glitch of some form. Just for fun, and to make the idea clear, let's have this process available over the network. Let's call this thing "Splork".
On the left we have Linux, on the right, we have NT. Both NT and Linux are outfitted with user-mode processes which will trigger "Splork" on the opposing machine... lets's call this process "Bang".
Bang fires off Splork, if the target machine is still alive, Splork calls Bang and hits the the other machine... Eventually, one machine will stop shooting back.
Equip a third machine with a couple reset switches, and some monitoring software....
Let this run long enough to generate some good sample data... and see which machine is more tolerant of being Splorked by bad hardware.
Hey, it is no more strange than watching a machine lock up for no good reason... Don't rip this idea apart unless you really see a flaw in the logic.
Looks like the Linux world is starting to wake up the MS giant. We kicked their ass then they fixed stuff. All then have to do now is fix the crashes and their OS may survive. If it stops crashing and starts working then it will go somewhere. I will still use linux though.... More configurable.
The above is not worth reading.
At the end of the day, the smart companies have only two questions about IS technology:
1: Can I do more with this?
2: Can I do the same job cheaper with this?
All the other numbers are indirect data, trash talk. Management--especially smart management, doesn't directly care about MIPS, MTBF, or benchmark numbers. They care about the two questions above, and care about the other numbers indirectly because those numbers tend to be good predictors of the answers to the real questions. In this business, when almost everything is potential, these early indicators are very important, because you can't get good answers to the top two questions.
You have the same thing in sports. You can measure free-throw percentage, height, weight, slugging averate, save percentage, and a host of other details. But at the end of the day, only one question matters into it: How often do you win? All the rest are trash-talk numbers--good predictors, but not the bottom line.
In sports and business, you have to have those trash-talk numbers for people to give you a chance. If you weigh a trim 175lb, nobody in their right mind is going to make a nose tackle out of you--you won't get the chance to show the coach that you can topple the 325lb center. If a product has enough benchmarks damning it, the vendor will pull support and recoup its losses.
This is why Linux can ignore the trash-talk and go straight to increasing capabilities and lowering costs. Linux isn't a business; vendors cannot cut all support. Nobody has the power to tell Linux that it cannot enter the IS world. It can't get cut, and can only get discontinued if every Linux geek in Creation decides to spontaneously drop it. Red Hat and Caldera can go belly-up, Torvalds and Cox could be swallowed up in earthquakes, and Linux will keep on existing.
So long as Linux exists, it can win. With the development advantages it has, it can win well. It needs a foothold in some IS shops; it's getting that, or has already gotten that.
If Linux wins, it is going to start by revolutionizing an IS department. Some big gun like AOL will see the potential and let it start taking over the infrastructure. It will work. Forget the runtime, forget the performance, it will do the job for cheaper. In the business world, such success gets copied. People look at the company that pulls this off, ask how they do it, and see a room full of Linux boxen.
The IT budget will convince more smart managers than any amount of benchmarking will.
PR is still relevant, but only in the short term. Good or bad PR can accelerate or slow the rate of Linux installation. In the long term, however, the success of Linux will have nothing to do with the benchmark numbers and have everything to do with the budget numbers. If Linux can do the job cheaper, it will win. If it can't, it will remain a hobby OS.
But the good news is that, unlike a corporate product, short term effects cannot destroy the long term picture. Linux will have all the time it needs to fit into the corporate structure to its best abilities.
--The basis of all love is respect
> And no more excuses. Linux is not the fastest. Deal with it.
;-)
khmmmmmmmm; two points:
* I do some tests from time to time; one of my jobs is to select hardware/ software for our university labs; I remember my disappointment when I realized linux kernels 1.2.* are slower than 1.0.* (in some tasks); this was due to some bugfixing
* did I missed the software configuration description of the tests? after the Mindcraft tests I did some tests for myself (we do have Dell with 4xXeon400 and 2GB of RAM running Linux); I optimized Apache for large load (1000 of subservers running paralelly, large file limit, etc.); this seemed to work fine even when the number of simultanously running sessions exceeded 2000; I can be wrong, I had no time to make those tests properly and documented; but I would like to see their setups...
best,
robert
Go read the O'Reilly paper again before you resort to obscenities. Yes, you can get a server from a workstation by changing two keys, but doing so causes the system to set a number of other performance-related keys differently, as the original poster stated and the paper would confirm, if you actually bothered to read it.
Wow... It's taking Microsoft, 4CPU's, 4Gb of RAM, high performance RAID controllers, quad NICs and a team of MS's top software engineers dedicated to tuning the system to beat Linux.
:)
Who in real life will have this kind of hardware and get that kind of support from Microsoft?
Boy are Microsoft in trouble!
Deleted
Come one now. This test wasnt any better then the first one. There are very few facts in it. They didnt even say WHO set up the Redhat box's. For all we know it could have been a 18 year old Redhat PHONE TECH!.
How many hits did the Linux server and NT server drop?
Did any of the servers crash?
What Kernel/Server versions where they running?
Can we see the config files?
This test is WORTHLESS and is only good for making FUD. Maybe NT can realy beat Linux..OK. But they could at least give out some info on what Linux realy fell short on so it can be fixed.
And again, version #'s would also be nice. They could have been running kernel 2.0.30/Redhat 5.0 and some old ass version of SAMBA. FOR ALL WE KNOW!
I have to return some videotapes...
I personally find it amazing that MS people are out cheering in the streets, this was a very weak and narrow victory for NT, on the single processor NT was only 50% faster and on the quad system less than 2.5 times faster. This for an OS built and backed by a huge corporation that has pumped billions of dollars into the creation of NT.
Let us all keep in mind that it was essentially MS that set the conditions of this test. The hardware was built for use with NT and was further tweaked for use for NT by using 4 ethernet cards. A neat trick but still a trick favoring NT. The quad processor test was very eye opening, for a price of 10 to 15 times that of a single processor server, NT's speed barely increased over 2 times. This is hardly a selling point for quad proceesor machines or for NT for that matter.
On the web server tests, an essential flaw was found with the IP stack on Linux. It is said that this problem has already been remedied in the 2.3 kernel. Also, Appache is being redesigned with a static page engine. Considering, the hardware used, the nature of the static page test designed to favor NT, and the above mentioned fixes MS NT's web serving speed advantage will be a very shortlived phenomena.
On the file server tests, NT also outperformed Samba underwhelmingly. Let us keep in mind that file serving is run on the SMB protocol - an MS protocol that not only MS controls but that they are less than helpful about when releasing technical info on it. Samba is entirely reverse engineered. In addition, they needed to run it off of 4 partitons to enable it to beat Samba. While the 4 ethernet card set up and the static page test make one gape because of a total lack of real world usage, I can not believe that ANYWHERE in the world that ANY NT file server was set up with 4 partitions prior to this test.
In addition SMP has been in the Linux kernel for less than six months and has hardly been run on such a high end system but was only outperformed by NT by roughly 2.25 times. Also weak points and bottlenecks were identified in the SMP abilities of the Linux kernel. These are likely to be ironed out very quickly.
So NT beat Linux not very soundly on hardware built for NT on a hardware configuration designed to give NT a performance edge. Congrats to all of you NT people out there. However, instead of enjoying your victory you should be wondering how this "hobbyists' OS" came so close to NT that has billions of dollars sunk into it on a test clearly designed to favor NT. In addition, with the knowledge the Linux community has garnered from these tests, it will obviously not be so long before Linux can beat NT even on even such unreal tests as these.
By raising such tests, MS has put Linux in the limelight and has given it credibility that it has never had before. In addition it has let us know how to improve Linux. Thank you MS for the help and if you think about it just a little don't you think it would have just been better not to have said anything at all?
>If you're just talking software, dollars/*anything* linux is zero and will win.
Not necessarily. Someone may be evaluating a supported distro, such as RH6 boxed, in which case there is a price. However, unless you install via ftp or borrow a cd, there is always a price. Its just that for Linux the price is orders of magnitude lower than MS. But yes, I was just referring to price, since hardware is the same no matter what OS is installed (in this case), and Windows and Linux run on the same platforms.
UnixWare is supposed to blow NT out of the water on systems with 4 or more processors - it has very linear scalability.
Anybody tried it on these sort of tests?
Bob
>> but if the main process dies then all child threads die.
>AHEM, HELLO, BULLSHIT. This exactly DOES NOT
>happen under the windows process/threading model,
>and IS what happens under most other models. I
>personally consider it a FEATURE when the main
>thread cleans up the suboordinate threads when it
>dies. If the main thread dies, YOUR APPLICATION
>HAS QUIT. There is no reason to keep the rest
>of the threads around.
How does "cleaning up" differ from dying in the context of my statement? The child threads are still killed off right? All those users out there who were being served by the child processes still lose thier services... right?
>I defy you to show how EITHER model adversely
>affects system stability.
Stability is a loosely defined metric. If stability is considered to be defined as the ability to continue doing it's appointed task then the forking process model has obvious benefits; however, if stability is defined as the ability to remain in operation after a server/daemon failure then then both models have benefits. NT's threaded model cleans up by killing all the child processes and then the server must be restarted - forked process models just spawn a fresh daemon automatically and continue as if nothing had happened.
If the server's purpose is to serve then it should do it's job without fail. If a model can be found that ensures and defines stability as the ability to do it's appointed task within it's operating parameters then... that is the model to choose.
At the present time, it has been pointed out that Linux on a single processor PII class machine equipped with 256MB RAM is capable of maxing out a T1 connection and is still stable enough to take a pounding while running multiple services. I'd call that a realistic expectation from a machine designed to do a specific job.
Codifex Maximus ~ In search of... a shorter sig.
that's not really a decent comparison, because the win32 port of apache was only recently completed and is still not considered stable.
also apache is built around unix and unix like systems, so architectural decisions made to improve performance on unix like systems mean it takes a hit in a lot of areas on win32 platforms.
that's really no more fair to windows nt in benchmarks than it would be to use iis on linux (assuming it existed), and most people here would probably start a small riot if they tried to do the tests that way.
If I don't put anything here, will anyone recognize me anymore?
not in Linux case, fork() is just a special case of clone().
>Especially when you have to start communicating across processes.
Not the case to serve static pages, right ? Not to mention IPCs aren't for free between threads either (locking).
>Threading isn't just a nasty hack to get speed, it's there cause it's a good design - why do you think a mutithreaded ip stack is going to be incorporated into linux soon?
What's the difference between processes and threads ? Share of memory space, no more, no less. And no better, no worse. Linux IP stack won't get "multithreaded" as you say, it will be de-serialized. And if you don't see the difference, get a book on "standard" (not microkernel one) Unix architecture.
>And I wasn't saying that Linux IS guilty, I was saying that it's more likely linux would be guilty of leaving requests idle than what you said about NT (eg. don't make stupid claims about NT).
Same thing about Linux, I don't think the choice of processes/threads influence the equitability of the scheduling more than the scheduler design (hint). NT and Linux have different schedulers, but I don't think either is significantly "unequitable", right.
I am surprised by the reactions from the Linux people on the board. I feel compelled to play devils advocate and point out that Linux has been beaten on 3 proffesional tests. However the only reactions I have seen are storys about how the had one seen NT crash or that "in their experiance" NT was slow. This may very well be true, but benchmarks do not bear it out. If Linux is to compete for mindshare it has to be faster on the benchmarks! Maybe their was something wrong in the tests, but in the end, people listen to these big name mags. ZD may not be the most trustworthy (I'll never forget how they said that the i740 was faster than a Voodoo 2 and that a Rage Pro was faster than a Riva 128) but people listen to them. Instead of complaining about how the benchmark is bad, try looking to see if their are actually problems there. And remember, PC Mag WAS trying to make an effort to be fair, they wanted to come out of this looking like the objective king, so Linux did have good support. Also, they did point out cases where the Linux people said package XYZ would change the whole benchmark but it didn't. Lastly, even if NT is 50% faster, if your job isn't mission critical (and you don't trust that to NT OR Linux) then 50% is a lot. So is 200% I bought a Voodoo 2 even though it was ONLY 70% faster than my old card, but it made a differance.
A deep unwavering belief is a sure sign you're missing something...
Not to mention NT has way to go before scaling against *all* commercial Unixen (many benchmarks around, pick yours)...
The fact remains, if the free operating systems are so horribly bad, why is it so many private users and corporations that use them FEEL like they are more stable and better performing? Oh well, just my righteous two cents. :( ).
Because it doesn't match our day-to-day experience ?
Since I moved our accounting database from local box (PII NT workstation) to the server (P233MMX Linux), the guy on the NT workstation said he observed an *increased* speed. And yes, the files are shared thru Samba (I know, some client-server architecture would be better, but hey, the stuff is not exactly OSS
Agreed, that's a small LAN with few clients (mix of NT workstation and Win 9x boxes), so even if I believe the benchmark figures are for real, I can't reproduce the tests.
But I can say that in my real life case, Linux is a much better solution that NT. I feel like this benchmark is totally irrelevant because it doesn't grab the cases when Linux is used (at least right now). It mostly shows that NT features that Linux lacks, like 0-copy networking, are nice to have sometimes, that is mostly during benchmarks.
How can we fix it if we don't know whats wrong? They gave out very little info about the 2 setups. For all we know they are running a NT verson made JUST for this test VS. Redhat 5 and some old ass version of SAMBA.
I have to return some videotapes...
Quote from my previous reply:
Personally, I don't really beleive its a cluster,
However its not just the disks.
It has 4 ethernet cards, each bound exclusively to one of 4 CPUS, and no doubt each ethernet card is accessing files on a different HD.
It isn't a cluster, but its a dammed expensive, and I'm sure unusual setup.
It doesn't crash daily. That's Linux FUD.
:p)
NT4sp3 *NEVER* bluescreened me after runningit for a year on my desktop.
I don't have to sit in front of an NT box anywhere anymore. But at my last job, about a year ago, I had a dell ppro 200 with 64 meg. nt4, sp3. And I'll agree, it didn't blue screen often at all. Also, I didn't use this machine very heavily. IE 3, 2-3 3270 emulators, and lotus notes. You could watch my memory usage in performance monitor and watch my free memory drop by 5-10 meg each day(I only shut down once a week, who wants to wait for an NT box to boot in the morning?). By a little over a week, it'd freeze(no bsod though
"Any software which causes Windows NT to blue screen is a bug in Windows NT." -- Steve Ballimer at the release of WinNT 3.51
--
Ben Kosse
Remember Ed Curry!
i started this..so i'll end it. i've had crashes with NT sunder the following SP's :
SP1, SP4, SP5.
i've not tried SP2/3.
I wish I could rember where I read this but it was about a year ago. I rember reading that MS products are designed to beat any comptitor in bench mark test but with the price of security. So OK if NT is twice as fast as Linux in web hosting and SMB network files. But the real question is how long will it work for, a day, a week, 2.5 weeks (My best knowtestable uptime for NT). Compared to uptimes in Months, Years, (I cant say decades because Linux is only 8 years old). I have done different benchmarks comparing a linux system to Win 95 and NT and in my test Linux out performed NT by a factor of 10, these are my rough results
A simple test to count and display 1 to 1,000,000 to the screen use a basic for loop and register INT
Duel Pentum 200
Linux text mode (2.0)kernel 58sec
Linux in Xmode (2.0)kernel 2min 30sec
Windows 95 Graphical mode 10 Minutes
Windows 95 in dos mode 30 minutes
NT 4.0 Graphical 7.5 Minutes
NT 4.0 Full Screen text 27 Minutes
Well I say in this benchmark test Linux won. But I could of had flaws I used GCC for Linux and Borland Builder for the Win32. I should of had 50 xterms open to simulate the system load of nt (I did that and got around a simular time). I should of taken the fact when Linux displayed the numbers to the screen The screen flicked every couple of sec while NT and 95 the display moved smothly and no flickers or small pauses.
Just to proove my point the benchmarks are the proof of the better system.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
Actually, Dell already replaced the motherboard in this server after we had installation trouble. I don't think hardware is the issue - Dell has checked this out. Besides, it's brand new. If any computer should be giving us hardware problems it should be the old throw-away box running our Linux web server.
As for investigating the problem, we just did an out-of-the-box install, nothing special. We don't really have time where I work to investigate too deeply, but that's another issue.
Our Linux box was a plain vanilla, out-of-the-box installation as well, and it has given us no problems at all. I basically set it up and it runs with no intervention on my part.
If you're suggesting I'm not putting enough effort into system administration of the NT box relative to the Linux box then that shatters the stupid assertion that NT is easier to administer, doesn't it.
BTW, I'd never work for an anonymous coward like you.
I have been running W2K Professional Beta 3 since it came out in March on both my home and desktop workstations. It has not crashed. It is faster than NT4 SP5. It configured everything in my box perfectly without intervention, except the ZIP drive. That requires a registry hack to enable. That I believe has been fixed. It has run every app I have thrown at it, except games. Any game that ran on NT4, like Half-Life, runs on 2K. The NVIDIA TNT driver is not mature enough yet to run stuff like Star Wars Episode 1: Racer, but the Beta 1.88 NVIDIA driver is starting to show some promise. If MS continues down this path, W2K will be a great OS for the desktop. Haven't done enough on Server to comment. I thank the Linux community for the state of W2K. Without the challenge that Linux was the best and Windows NT, Microsoft's bread and butter for the next 10 years, sucked at everything, I doubt MS would so vigorously be persuing stability and reliability. Long live competition, arrogance, and fear to deliver the best products into consumers hands!
I think that they should do the test with a freebsd server and samba/apache. Freebsd is much faster in response time than linux. I was also thinking, isnt apache ported to winblows? Why not try using apache on a NT server?
Ok, we see where the weeknesses are in Linux, and we know what to do about it to make them go away.
DONT SIT THERE COMPLAINING ABOUT LOOSING! FIX IT GODDAMNIT!
Nobody likes a whiner least of all the corperations or the general public. It doesn't matter if the tests were rigged in favor of M$ or not. we know where our weeknesses are and we can fix them. if the results are wrong or padded or something then fixing the weaknesses with just make linux that much better than NT or any other OS.
denial or even functional denial does not serve a perpose.
If you want a system to work then make it work and fix the bugs too! don't point fingers and say. "He's lied!" or "The test was rigged!" or even "They were paid off!"
take a look at your failures and fix them!
If at first you don't feel good.... suffer like the rest of us.
How about not for an internet site? One of the tests was for a file server, which is usually not on the internet... throwing them both up "out of the box" would be a good test... how many people really spend a lot of time tuning their NT box? Not many.. they throw it up, and let it run (and BSOD)...
That would be a good test, IMHO...
Are you updating your NT server with SP without reboot? Service Packs released with less than a year period between them. How can you keep your NT up continuously for a year still following service packs?
What have you been done to track the reason for Linux instability. To have server crashing every 5 days is a bit annoying IMHO, so you'd have something to be done with it. Many people said what Linux is stable for them, so maybe it is the problem with your setup?
FreeBSD has its problems too. Read the freebsd-stable mailing list and see what kind of problems FreeBSD users have. There is no perfect OS!
Exchange server is not a joke. It's the saddest story any software company ever delivered.
You shouldn't be able to crash NT from userspace if the box is correctly set up. Problems are:
1) Too many settings and file areas have too unrestricted access level by default on NT. Try restricting access according to recommended settings.
2) NT users have Win9x at home. They demand the same access level at work, and "helpful" admins cheerfully add users to the WKS' administrator group. You would expect more Unix crashes if everybody had root access all the time, too, eh?
I really could care less if you believe me or not. I have produced and re-produced these results over and over for the last 8 months. If it were NT that was actually more stable, I would be praising NT. But that is not the case.
NT is faster than netware -- look at (who else?) mindcrafts benchmarks. NT is also faster than a cray SV1 BTW...
i've written multithreaded stuff on OS/9..its crude and primitive but it does give you real time performance out of the box. its kinda like CPM + UNIX combo...
Without going into details that I should not go into, we do not develop new hardware, but we do develop solutions using existing products/technologies. Products know to work with NT, just put into the mix with other existing products. So the fact that NT crashes under heavy load when Linux does not shows the stability of the OSs using existing technologies.
SP4 and we are revising our test plans to include SP5. I have not started reliability testing with SP5 yet.
I dont know what you have been smoking here but win gui, as a whole sucks. It is my impression that you have never used any real x windows system other than the default one that comes with linux. I have been running X for many years now and never once had it crash. I have however had nt crash on me and i know that nt could not have even 1/100 of the uptime of a linux box could.
its cost. theres no one out there to pay for the tests to be run. would you like to contribute ?
You're assuming we get the famed BSOD. We don't. We get what looks like a normal GUI desktop, but everything is locked up. This has even happened overnight with no users logged on. Perhaps directory replication was taking place with the PDC. But if this is so it casts serious doubt about NT's place in the "enterprise" - a big network will have several servers and the normal housekeeping shouldn't cause problems. Also, directory replication is a built in part of NT, not a third-party add-on. Shouldn't it work flawlessly?
And regarding your story about a third-party software package, this highlights one of the problems with NT: poor separation between kernel space and user space. If you can't run "third party" software for fear of causing system instability you've got a pretty useless OS in my view.
By contrast, I've been using Linux for a little over a year now and I've never seen a stable kernel crash. I've beat on it pretty hard, too. I do alot of programming so I'm always debugging software, much of which uses pointers extensively. I've built several beta software packages. I've reconfigured and recompiled my own kernel several times. The worse I've ever experienced with Linux is having to log on to another terminal to kill a runaway process - never a crash.
As for my "pretending" to be an NT administrator, that's just a silly statement. It's a job position and I currently fill it. To be perfectly honest, though, I'm only self-taught as a network administrator, mostly OJT. I'm a programmer by education (Comp Sci) and a pilot by training (USAF). I am studying, though and am pursuing an MCSE.
But then, I thought NT was supposed to be so easy to administer?
Well, I've read a lot of comments here and I think it boils down to a simple desire. If speed is what you're looking for, maybe right now NT is the solution. The trade-off is that you sacrifice reliability and stability. Take race cars for instance. They tune up for qualifing because they can get more speed, but at the same time the mechanics know that tuned up settings will never get them thru say, a 500 mile race. Indianapolis is a great example of that, before the IRL came around, they could push upper 220 MPH laps, which you'd never see in a race until the very end with only a few miles to go; when speed is everything. To last 500 miles, they tune the cars back for reliability. Linux I would say is very similar in that respect. I would think that many people in the Linux community could easilly accept this gracefully, knowing full well that the Linux/Apache combination may be slower, but at the same time will provide much less unscheduled downtime. M$, I think, would rather have great performance bechmark tests and inferior stability, then the opposite because benchmark tests are what get published and people that aren't that well informed about computers will read that. It's like when your new to the web, flashly graphics may awe you, but after awhile, when information is what you're looking for, text pages do just fine. M$ strives for looking good, people that know computers know that looking good doesn't always get the job done. If reliablility is what you want there are far better solutions than NT. For example,we have an SGI Octane at work that acts as a file server and an intranet server, I have no idea when the last it went down abnormally was; maybe never in the 2 years its served that role. Could you say that about NT, I highly doubt it. If NT had that reliability and stability I think that the people who post their concerns here would have a reason to really be worried, I personally don't see the need to pout over NT's speed because Linux is still VERY young, (especially in the commercial world) and this is an obsticle that I would assume be easily tackled in the near future. Hope I'm right, have a good day.
This study has just proven to me that linux is more than fast enough. PC week said the linux box put out 1800 requests/sec to 3500 for NT, however after I did some calculations I realized that there numbers are way higher than they need be. 1800 requests/sec comes out to over 150 million requests / day. I think I can count the number of sites on the internet that need that kind of static web serving on my right hand. Oh yeah and then only 13.4 terabytes / day of data. Hmm. I really think this test shows that they both put and more than enough. So now we need tests showing stability. Let's see what the actual output is for a week, or even a month at these speeds, I am dying to know.
javac
Whilst the test have highlighted some 'limitations' of Linux in a lab, and whilst this looks bad for Linux, and is something that we the Open Source Community should address seriously if we wish to gain widespread acceptance in the business world, it really fails to represent reality. My company is using a Dual P90 as it's main webserver, not a quad P500. If we got 1000 hits a second on our webpages we would probably die of heart failure. I saw a report that suggested that even the Charles Shwab e-trade site only gets aprox. 642 hits a second. Looking at it one way, the test showed that Linux for $0 is perfectly capable of performing the same job as NT for $1500 (once you add in mail etc..). Whilst NT might be faster in _HUGE_ volume tests, in the real world, there is probably little or no difference except in the area of stability.
However... for very small organisations, I run an ftp server, web server internal DNS, NIS, SMB (there is one Win95 machine) etc... for a small network comprised almost entirely of old 486s w/16MB mem, and 400MB HDs, Linux is the _only_ choice. NT wouldnt run on these machines and 95 isnt pretty! My home LAN which cost me less than $300 for 6 machines hub cables and all, plus $1600 for my main system (which I bought new and is now rather an outdated P166 w/8.4GHD+3.5GHD 64MB RAM) (7 machines in total). It performs great for my needs. If I were a small business, I think that I would have to think twice or three times before outlaying large sums of money to M$ for a system that was so over my needs, instead of using a system that would cost me so little.
Linux offers computing 'solutions' where NT offers computing 'problems'.
Bang/$, Linux will always win. Cost of upgrade since Linux 1.x software and all... $0
Cost of upgrade since DOS for M$ softward and all?...
anyone care to speculate?
Do we need to improve Linux' high end performance just for the sake of benchmarks? possibly not, but It wouldnt hurt.
Everyone is living in a personal delusion, just some are more delusional than others.
Yes, I know. Lets see, NT does not have any of the Microkernel's alleged stability, so to argue that a microkernel is inherently more stable and then site NT as an example is rather dumb. Mac OS X is only out in a server edition and looks OK but hasn't proven itself. BeOS is single user only and still beta though quite impressive. Show me a serious operating system that A) is a microkernel and B) has superior stability as a result.
So who has any real track record with microkenels? Microsoft. Hmmmmm....
Understand I don't mean any direspect to you, BeOS or Mac OS X but there's a whole lot more out there than just the desktop and right now microkernels are not necessary or worth the porting effort just to get enterprise stability. Linux, *BSD, AIX, Solaris, VMS, Tru64 Unix, HP/UX, OS400, OS390, they're all rock stable; NT, a microkernel, is good 23x6. Thats not much of a track record.
I browse with my threshold at 2 so I can't read my own comments :-)
The test proves that a Linux box talking to NT
clients. Is slow. Hmm, and the point is?????
How about testing Linux boxes talking to Linux boxes? It like trying to talk to little kids
about rocket science.
Regardless it is outstanding results for something
put together in peoples spare time, for FREE!
Microsoft has an entire empire to make their stuff.
Actually, I took a look at the memory dump, and this definitely is a video driver issue. I installed an unsigned NT4 driver, and it just took its time biting me on the ass. That makes this a "stupid user" problem. This will teach me to post on slashdot while drunk.
--Shoeboy
Your company would FIRE you if you ran a browser on a server machine? Would you mind disclosing where you work, so I can take steps to ensure that neither me nor any of my friends ever, ever go to work there?
-Graham
I don't know if this helps, but I've been able to trace a couple "solid lock" NT problems to SCSI cabling problems. One of these was on a new Dell server that shipped with a loose cable. NT doesn't seem to handle SCSI issues very well.
--
Business. Numbers. Money. People. Computer World.
IIS has a cost (as in WinNT). :P
You're maybe not saving license cost, but a) it has been shown that you need 5 NT admins against 1 Unix admin (given the same amount of server boxes - a Unix admin will do the diskless workstations for free), and b) running NT on servers will probably do the same thing to Total Cost of Ownership as running Windows on clients does: it goes up through the roof.
c't magazin has also tested Linux vs. NT. 4 CPUs, 2 Gig Mem, 1 NIC 100 MBit, RAID 5 (not RAID 1 as used @Mindcraft). SuSE 6.1
When serving 1 Million different pages with 4k each, Linux was up to 10 times faster than NT (30 pages/sec vs. 274 and more...)
They also tested with 4 NICs with 100MBit. And won...
Uups. Should be RAID 0, not RAID 1 :(
btw: they needed RAID 5, because 1 disks crashed on one day.. but runs fine the next one...
that I can find a configuration of machine that Linux can beat NT on if I want to.
Boss: What do the numbers say?
Accountant: What do you WANT them to say?
Codifex Maximus ~ In search of... a shorter sig.
I've worked with the same people and the same benchmarks and your claim makes no sense. In the case of NetBench and WebBench there is no code running on the server; it's just file requests. How can that possibly be Intel-biased?
They say the OS was the problem and not Apache or Samba so how about giving FreeBSD or some other BSD a try. Linux isn't the only open source operating system out there. On the bright side it is good news for Apache. It is as fast as Zeus.
Regards.
Ken J.
Linux vs M$ is not the first heated battle over performance and it will not be the last. One thing is for sure in this industry, the performance leader today will not be the performance leader tomorrow. I will personally have to endure days of the "I told you so-ers" in the near future. Does it matter? Not really. I still have an elephant to eat. And what is the best way to do that...one bite (err server) at a time. There is a lot more to selecting a platform(s) for developing/running large enterprise operations than simple benchmarks.
Keep in mind folks, we are the technical brains that keep corporate America running. With the very obvious shortage of experienced talent out there I dare say we will always have the upper hand when it comes to selecting who we work for and what platforms we do that work on. To that end, if management starts trying to define implementation based entirely on benchmarks...well I hope they have been taking some night classes.
**** Sworn to Fun, Loyal to None. ****
It's so nice of Microsoft to pay for this apache advertising. Just as a point of reference, 1800 hits/sec is the same as 155,520,000 hits/day. I think that it's safe to say that noone in the world gets more than 150 million hits per day of static content. Wait, there's a better way:
1800 hits/sec * average 2k/hit * 8192 kbits/kbyte = 29,491,200 bits/sec, or 29.5 MBits/sec. What's that now, a T3 line? I know that a T1 line is 1.5 MBits/sec. Ok, so apache on one of these boxes can fille the equivalent of 19.6 T1 lines by itself. If (a bit more realistically, how many 2k files get those types of hits) those are just 10k files (let's not get into pictures), that's 147.5 MBits/sec, more than filling a T3 line, IIRC, and definitely filling up aapr. 98.3 T1 lines.
What's the problem with Linux/Apache, now?
May I suggest, if you can afford this sort of bandwidth, that you buy one of those 32CPU sun E10000 servers and call it a day? (or a server farm of linux boxes, since you're serving up static files.)
Oh, if you're serving up >1800 files per second of 2k files, who are you?
Oh, one more thing. If this is all on an intranet, you'll still need Gigabit ethernet if you're serving up the 10k+ files, so the sun box still applies to you.
They laughed at Einstein. They laughed at the Wright Brothers. But they also laughed at Bozo the Clown. -- C. Sagan
It doesn't matter how much of a pounding you give the OSS community. We'll ALWAYS be there for the NEXT round. Better, stronger, and faster than before.
Chas - The one, the only.
THANK GOD!!!
Chas - The one, the only.
THANK GOD!!!
If Linux's bottleneck was a non-multithreaded IP stack, wouldn't FreeBSD have similar problems? IIRC, the FreeBSD kernel is not multithreaded, so it would scale even worse than Linux on a multiprocessor server.
Does anyone have any FreeBSD web serving benchmarks?
cpeterso
OMG! You said something good about Windows and bad about Linux. That is punishible by death on Slashdot! I swear, some of you Linux people need to get a grip. I admit, I have never tried Linux (though hoping to pick up a copy tomorrow at a computer fair) but the way yall act is pathetic. "Linux is SOO much better than MS. EVERYTHING MS makes absolutly sucks. MS has NEVER made a good product." And if Linux ever gets bad score, like this case, then it's always "ZDNet fixed it so Linux would lose", or, "They should have tested it on a 386 with 2 megs of RAM, yea, then Linux would win" They forget that NO server even uses anything close to that bad. I'm a Windows user, and I like Microsoft products. If that makes me a rebel and an idiot in your eyes, then the heck with you!
linux lost, oh well. just remember that while half you people are crying about linux, and the other half saying that Open source stuff doesn't work, there is the shining example of what OSS can do sitting there serving up cdrom.com. I've yet to see Microsoft pull that kind of throughput and uptime (at that cost). But it just won't be thought of as much, because the mood still seems to be "if it's not Linux, it's crap!", and evidently, the *BSD stuff just isn't as cool. Whatever.
I think the real danger is that the community will feel pressured to "fix" the problems before the fixes are stable.
I really don't think that will be too much of a problem. After all, that's what the development kernels are for, anyway (and patches won't get ported back to 2.2.x until they're relatively stable). If a business insists on running a development kernel on a machine, let them reap the rewards as they may be.
Because those are IIS servers and you can't bring them down..
;).
Well, consider the fact that there are 96 Compaq Proliant 5500 (quad ppro boxes) running microsoft.com on a fiber optics backbone... Check it out for yourself. It must be nice to have the cash to fix software shortcomings with hardware
--- A Jesus Fish eating a Darwin Fish only proves Darwin's point.
NT probably COULD run Hotmail, it wouldj ust require a much larger investment of time and money than they want to put into it. microsoft.com runs 96 Compaq Proliants and has a good deal of files to throw around, Hotmail would be about the same size, so it would require another 96 Proliants for Hotmail, an investment I dont think M$ is going to make. AFAIK Hotmail uses FreeBSD not Solaris.
I'm a loner Dottie, a Rebel.
I realize that this comment will be moderated down for being off topic, but I feel the need to digress in response to the increasing criticizm of those who submit comments as "Anonymous Coward".
/. here and there during the day. I don't want to log in from every computer I use, simply because some of my coworkers also read /. from some of the same machines, and it's just too much trouble to log in and out all of the time.
/. comments that automatically come out with a score of 2 or greater. The algorithm would be something like this:
When I'm at work, I sometimes have a few minutes to read
I know that it isn't often the case that an AC's comment gets favorably moderated, but well, neither do most of the bold individuals with names who submit comments.
As for flames, they happen.
I'm thinking of writing a program to generate
// begin post
[mention obscure fact about issue]
[predict that post will inspire flames]
[express disregard for above prediction]
[take 6 or more words to express linux community solidarity]
[talk about how this topic relates to something at work]
[claim not to be surprised by results/information/story]
do one of the following {
[mention a famous OSS advocate]
[comment something to the effect that "linux will win"]
}
use several of the following {
[IMHO]
[AFAIK]
[M$]
}
do not use [LOL]
// end post
One of my CSE professors at the University of Washington worked with Dave Cutler at DEC. He said Cutler was the "best programmer in the world". He said Cutler would always have two piles of printouts on his desk: a very short stack of his code and a tall stack of all his test code.
Unfortunately, Cutler is now in his own little world, racing cars and working on a Win64 project (codename "Sundown" harhar) that will probably never stabilize..
cpeterso
We must look at this this way: The first tests MindCraft did returned very little data to help in fixing the problems and it was 80% NT marketing. This was bad for Linux because we got nothing from it. We were able to get a rematch where we get reams of data on what/how/why the results fell where they did. Linux still did pretty well when you consider it can run on a 4MB 486 router up to multi CPU servers in a cluster to rival super computers. We got tons of data from this and can only make things better. This wouldn't have been posible without this test. We should thank MindCraft for working with that Linux group and get back to fixing/testing Linux.
I don't think Microsoft can scream as loud as it wanted the first time the test were run. We win on these two issues. IMO.
"Anyone who stands out in the middle of a road looks like roadkill to me." --Linus
> How does Linux compare to Mac OS X?
There haven't been many comparisons yet of Mac OS X's overall capabilities to the other major OSs, partially because OS X is still in its infancy and partly because there's still some lingering doubt over Apple's long-term plans for the system (the scientific term is Coplandophobia, a fear of dead-end business decisions spurred on by a lack of a clear strategy; not fatal if caught in time).
It appears that it will improve (eventually) over AppleShare IP for file/web sharing, although granted that's a long way from replacing NT as your server of choice.
I know, I know, MacWorld looked at it in their July issue. However, IMHO that review shouldn't count for much; I find it troubling that they had to compare it with a multiprocessor IIS/NT box because a machine comparable to the G3/400 used in testing wasn't "available at press time". Sigh. Not that one necessarily expects much from the magazine that over-trumpeted its exclusive coverage of Apple's purchase of Be, but why bother testing it if you're going to stack the odds like this? It would be like testing X Server against netBSD on an SE/30; they both work great, but one is obviously going to outpower the other.
Hopefully, by the time major publications and organizations get around to looking at OS X and/or Darwin as an option along with the other BSDs or Linux or even NT, the hardware will have caught up to the promise of the software (viva la AltiVec, baby, yeah!) and Apple will finally be a viable alternative to Windows in the higher-end again.
Please, no flames about the netBSD crack; I'm shopping for an SE/30 on eBay even as you read this for the sole purpose of proving to my nonbeliever friends that a 10-year-old computer smaller than a 13" monitor can still be useful to society.
I use Macs for work, Linux for education, and Windows for cardplaying.
This is now the chance for Linux to show its true strength - the Open Source Community. Even in the midst of the Mindcraft controversy, the kernel and Apache developers were busy using the Mindcraft tests to identify and fix the real performance bottlenecks that Linux has. Don't worry, we'll catch up.
Meldroc, Waster of Electrons
All of your stuff is completely relative.
Price/Performance
It's all related to how much you pay your admins and how well they administer your system. This isn't a function of the OS. Yes, Linux costs less out of the box, but an NT admin is going to have a harder time (and thus charge more) to set it up than he is an NT system. If a business currently has functioning NT systems and competent NT sysadmins, why should they switch to Linux?
Clustering
How many small businesses who are choosing between Linux and NT need to, want to, or care about the ability to cluster? People who care about this benchmark are not the same people who need to run clusters.
Other Hardware Configurations
How much would it cost for a company to build a Linux-happy system? Most systems built today (and the systems that we want Linux to run on) are built for Microsoft. You'd need a custom-built, custom-designed solution to truly grab all of Linux's power, and that costs money, either in man-hours or purchasing power. The results of this test would've been far more atypical if they had built both machines finely tuned for Linux. At least this time around, they weren't blatantly geared towards Microsoft.
Security
Security, I'd say, is 75% system administration and 25% OS. Linux has its security problems as well, most of which can be plugged up with effective network management. Many of NT's can, too. MS may be a lot more apathetic to security concerns, but they don't run the systems, they sell them. I don't consider Linux or NT any more secure than the other.
Stability
Stability can be completely a function of management. I've heard stories of Linux systems stay up for months or years. Guess what, I've heard the same stories about NT as well. I've also heard stories about unstable Linux systems. I've seen no long term studies done on system stability, so everything I hear about stability I file away under anecdotal evidence, not hard verifiable data.
Change real world needs
What's good for the goose is good for the gander. I don't see how this benefits Linux. Change the system and, whoa, Linux might perform worse under that setup. It happens to both types of OSes, and before you say, "It happens to Linux less!" find some hard data, not stories.
The Future
Past trends do not determine future performance. I doubt Linux will keep up its 212%/year growth and Linus has already said that upgrades aren't going to be as drastic as 2.0 to 2.2. Don't assume that Linux will advance in the next three years as it has in the past three years.
"I am a linux(still pronouncing it lie-nucks...is this wrong?)"
I pronounce it the same way.. and I care very little if it is incorrect. It sounds right, feels right, and it fits.
:)
Heck, even server-level SDRAM (ECC, CAS2) is going for ~$1/mb.
The command line will always be there, it just may not be part of the standard install or on the same CD.
I believe Apple wants to be sure (for good reason) that developers don't use the command line as a crutch and force newbies to 'tar|gunzip|make|make install' stuff.
- Scott
------
Scott Stevenson
Scott Stevenson
Tree House Ideas
Yes, Kurt, you hit the nail on the head.
/.'ers who could take Linux and Apache and merge Apache into kernel space, compromise everything else, and personally release an OS/Web Server combination that could easily beat NT in Webbench. The same goes for Linux/Samba. What they would have created is an O/S with dedicated application functionality.
There are probably a hundred (a thousand?)
Funny you should mention it. There are some kernel developers involved in writing khttpd, a kernel thread that serves static webpages. Last message I read about it on klm says that performance has recently matched apache.
Avi
At his pre-COMDEX address at Fermilab earlier this year, Linus addressed the issue of "mine's bigger than yours is".
What he said basically boiled down to this.
He doesn't care what runs "fastest". He wants it to run well, be stable, and generally hassle-free.
Chas - The one, the only.
THANK GOD!!!
Chas - The one, the only.
THANK GOD!!!
what we need to realize is that although uptil now linux has been designed for lower end machines(mainly cause the main developers and users till now could only afford cheap hardware), we ARE trying to move in to the enterprise, and there, the price of hardware, or for that matter, the software, hardly matters. even the regular upgrade cycle that M$ forces on them corporates is of no concern to them(after all the IT dept need to justify its annual budget!). So if linux has to make any inroads at all in the Corporate world, we need:
After all(and PLEASE, no flames, for once we should be ready to aceept the truth):
So what market are we aiming for? I think, THAT question really brings out THE central problem with the whole of linux. It is great for the folks who know what they r doing, and what they want (us!), and so it also targets us. But get out of the slashdot community, and the reasons for using Linux just start fizzling out. Th only real reasons u can tell to a non geek are that its free, and it performs better than windoze. Of course, RedHat Linux is not free(and to a geek, the option of downloading linux and installing it doesn't really appeal. frankly, even i, with my 4y+ experience with linux won't really try that), and these tests(which i am sure M$ is really going to tout al around) would show that its performace isn't really that great. Why should (s)he change, and give up most o' his/her favorite apps in the bargain? As of today, linux does not even have a stable browser, which can do java well. I hate Windoze, but i have to go to a Windoze machine(i don;t have any microsloth thing within 100 meters of my office) if i want to check out a java-loaded page. Want to quickly develop a GUI Application? A person who does not care about the OS, is definitely gonna chose VisualBasic over EMACS/GCC/Qt.
All OSes have specific aims:
BE: Multimedia...does great at it. no body even compares Windoze to Be in this area.
WinXX: ease of use, integration. integration, to what ever levels M$ carries it, does have a problem, but most people would take a frequent reboot, rather than give up the tight integration.
Solaris: Server, computing
IRIX: graphics, scientific computing
Linux: ?????
Where Is Linux? Where is it Going?
These are important questions, ANY movement must ask itself. If we find we cannot answer these, we would just turn out be Rebels without a cause. and linux would just end up being a note in OS history. Don't get me wrong...i work(about 8 hours a day at work, and another 4-8 hours/day at home) exclusively on IRIX and Linux, and have not even touched a WinXX machine in more than six months, i have been really thinking about the future of linux, and to me, at present, it doesn't look very bright. I would continue to use it, no matter what, but linux developers really need to do some kind of serious thinking about what they want from linux.
LinuxGhoul
Sigura Non Grata
I was (and still am) a little bummed that NT won. But I'm excited to see the rematch in 3 months, when a pumped & primed Linux, Apache & Samba combo wins with a KO in the first round. Microsoft is a beast, and it'll take 'em forever to update their server. They'll rest on their laurels for at least 18 months. We, on the other hand, will always be competitive.
I also want to see vanilla installs of the two competing operating systems... Commpared to when they both are tweaked... I wonder how much of the Window NT 4.0 server is really in from the distro CD or is it highly optimized code that can do only a couple of things really well and do the rest poorly?
Jason Wieland
EVERY company wants world domination. They are in business to make money and nothing else. And I be dammed if they aren't doing a great job at doing that. And if world domination, as you so well put it, achives even more money, than ANY business would want that. I bet if Linux was in everything, you would be quite happy, wouldn't you?
He is a M$ employee, or a contractor, ask him how much he gets paid for saying this. He knew how it was going to end up because it was all engineered. ;)
;)
Ok serious now. Why do you want to take this to a persona level? "LINUX IDIOTS"? Come on now, your mother HAD to have taught you better than that. If not mebbe you should have been beaten more as a child..
As for old hippies, I aint old. I still can't get my Linux box up proper (due to my own lack of knowledge, not the OS) but from the second I installed it I knew this (the win box) had numbered days.. Linux doesn't crash. Never has for me. Win crashes a *MINIMUM* of 4 times daily. And before it's said, it isn't 950a.
Let M$ have a field day. They'll be buried soon enough under their own loads of BS.
You fool! How can you give out the formula like that!? Now we'll have nothing but posters with high default scores! :)
Save the whales. Feed the hungry. Free the mallocs.
.
cpeterso
ok, so linux doesn't have support for it for the nic you mentioned... they should. I agree with the earlier post which questioned how realistic this was. A real world, real situation test needs to happen. One where issues about realistic situations are dealt with. I know what I experience. When ISP's switch to NT from linux, latency becomes a major problem. I've seen it several times. Not only that, but I say lets make it truely fair. get a team from microsoft and a team of linux techs together. Give them $5k to buy equiptment, and to pay full retail for software (i.e., microsoft not giving itself a price-break). Then see who wins. Linux is more stable, and can be more easily administered remotely. There is less of a constant maintanence cost. and exactly how reasonable -is- 1800 hits a day? as far as "linux" pronunciation is concerned, looong ago I had a sound clip that was supposed to be of Linus Torvalds pronouncing "linux." he isn't american, so don't say it the way an american would. its pronounced Lee-nux : )
vxfs (Veritas' file system) seems to fragment hardly at all, and actually HPFS (the OS/2 native file system) does quite well at resisting fragmentation, plus it has the handy 'extended attributes' built in to the file system for file type information, custom icons, etc.
If only HPFS has security, journaling and open source...
--
I would like a temperature reading on the processors durring the tests... I know that when I run my windows machiene and Linux machiene at idle I get a 5-10 degree temperature difference (In Linux's favor)...
jason wieland
I picked up a Damaged/Discontinued Compaq Pressario Pentium 166 for $300 (chip on the modem was burnt -- removed the modem -- computer boots).
I purchased Redhat 6.0 from Cheapbytes for $2.00+S/H.
I registered my domain (not infidels.org) for $70 for two years.
Other than the ISDN price, my entire domain/dns/mail/web/listserv-lite/workstation/etc cost me less than $400.
MS now sells Windows NT Services for UNIX for $150 to give NFS/Telnet capabilities on NT. Not counting the base price. Plus CAL's. No additional software.
I'd like to see a $400 server comparison.
A friend mentioned PGCC, which is much better than GCC from what I've seen, but I haven't been able to find it.. Anyone know where to get a copy?
Direct link to the file(s) preferred..
Posted by Jeremy Allison - Samba Team:
Actually, this is not the problem. The problem, as has been demonstrated, is the scalability of the Linux TCP stack with multiple processors.
When Samba is run on a Solaris system (which has a highly parallel TCP stack) on an x86 box then Samba beats all of the NT numbers listed for network throughput. The problem on the Solaris box is that the *file system* is poor, as this throughput is only obtained against a tmpfs (ram) disk.
When the Linux TCP stack is parallelized to the same extent Samba on Linux will beat NT in the same way, only the Linux filesystem is much faster, so no tricks such as tmpfs will have to be played.
Regards,
Jeremy Allison,
Samba Team.
Dunno if anyone else notices this, but it seems that alot of Linux-fans are poo-pooing the fact that the benchmarks were "run on systems that NT is designed to run better on" or "Run on systems designed for NT." Then they turn around in their next post and suggest running the benchmarks on a 386SX/25 with 4 megs. Umm, how isn't that doing exactly the same thing -- tailoring the hardware to a system Linux is designed to run on. Hell, we don't even support 386s.
Perhaps, while you are creating your new law, you would take the time to spell "Beowulf" correctly? I usually avoid spelling corrections (although lord that's difficult some days) but if you're going to call someone a "dumbass" and a "moron" then perhaps you should be concerned with how you appear as well.
This spelling flame contains no tyops :)
Your right to not believe: Americans United for Separation of Church and
It's easy to write fast software. Just don't do
any checking. I have written small, very fast
code. Unfortunately, if the user types something
wrong, it crashes. But boy is it fast!
It is no surprise to see an AC posting such an article. I think that it was just a couple of days ago when I saw some of the stuff that was posted on Mindcrafts site (linked to by /.) where people were bad-mouthing them because of the tests. It is posts like this that give the Linux community a bad name. It is one thing to call a foul, but it is another to act like a immature jerk and start bad mouthing a test that was conducted in a much more controlled environment. I think that ZD would have learned from Mindcrafts mistakes and corrected (most) of them. They are people too, and they DO make mistakes. Linux is not perfect, yes it is better than anything that has come out of Redmond...but it still has some maturation. Not unlike some of the people who post here. A true sign of maturity within the Linux Community would be to accept the loss _GRACEFULLY_ and go out and build a better OS. Then the next time that there is a test we can go out and show eveyone who is best instead of sitting on the sidelines pointing our fingers at the referees. Take the loss, no matter how hard that it is, and make what you feel so strongly about BETTER. Bitching about it ain't gonna build a better OS.
--------------------------------------------
Please give your mod points to others, Im at the cap. They will appreciate it more
... but we all know how the story ENDS.
The Micros~1 drones, PHB's and hunter killer slashdot troll-warbots are going to have a field day with this. Pay no heed.
support gun control: take guns from cops
This is the one thing that Microsoft would love to keep quiet, I'm sure. How relevant is two times the performance if you have to pay 10x as much to get it?
What's a copy of Windows NT Server run these days?
-Michael Pelletier.
Pretty much everybody has eschewed the microkernel model at this point, not just Linux. For a while HP, IBM, Sun and DEC were all paying lip service to microkernels and saying that they would one day move to microkernels, but none of them have. (Yes, I know Digital Unix is based on Mach, but its Mach kernel 2.5 not the Mach microkernel v3.?)
While microkernels may be more stable in theory, in practice good APIs and coding count for alot more.
I browse with my threshold at 2 so I can't read my own comments :-)
Although Linux doesn't have a PR agency, there's still the same kind of spin happening in this thread. Which (among the 2+ responses) is "MS has sacrificed stability/quality for performance". Lots of discussion about bypassing the HAL, super-secret internal MS interfaces, etc...
But we need to be careful. If you'll note, the article notes that Samba won in the earlier SMB tests tests because there was a performance hit in NT due to the transaction log. Which is a stability / robustness feature that Linux simply lacks, and would be better off having if availability and fault-tolerance are the primary design goal.
We're treading on dangerous ground... PR is like a game of chess, and the community needs to be careful about spouting out this kind of spin which can quickly become a rallying point and then proven foolish if it isn't well-though through.
This is all still using the old 2.0 kernel. One of the stips for the test was that no software released beyond the date of the original Mindcraft test be utilized.
So, a more scalable kernel is already out. Apache's had a couple revs. Keep hacking people.
Chas - The one, the only.
THANK GOD!!!
Chas - The one, the only.
THANK GOD!!!
There are some very important lessons to be learned here.
IMHO, this represents a huge step backwards in respect for the Linux crowd. After seeing the vitriolic comments about Mindcraft and anyone supporting MS, well, sadly, this tells managers that Linux advocates are simply ABM folks.
I'm not arguing ABM, but from an MIS standpoint, when you've got someone whith a huge, working MS investment, too many people have acted like children over the past few weeks.
And for those saying "we've got a fix in the works," let's not forget that MS has a fix in the works as well. I doubt it will be a cure-all, but certainly we should expect the bar to be raised.
(I can't remember my login!)
As pointed out in the article, what made Linux *slower* was the lack of a multithreaded IP stack, putting a ceiling on network throughput in this sort of context. I'd guess the CPUs were half-way idle the whole time Linux was losing... As someone said, it would be interesting to see how Apache and Samba performed under BSD... Anyway, this is on the way to being fixed.
>Some of you still feel the need to whine
Who's whining? By winning NT and Mindcraft still lost. You actually think this helped Mindcraft's reputation any? Nope. All it showed is that they are really only familar with Microsoft products, and doesn't seem to want to take the time and effort to learn about other products. Not a company to turn to if you have questions concering non-Microsoft products. As for NT the fact that this test proves that the numbers Microsoft were orginally claiming concerning NT's performance vs linux (remember their WWW page advertising this "fact") was pretty much bogus doesn't help their credibility all that much either.
You're economic argument doesn't "scale" beyond small business however.
Here's why -- Any company with more than a few hundred seats has a site licence contract with Microsoft. The cost is much more dependant on client seats then number of servers. This is to cover the client OSes and MS Office.
The cost of extending the contract to add a few additional NT servers to the mix is miniscule. Compare this to the cost of hiring capable Unix admins, and for any medium sized business, you're not saving any money with Linux.
--
Business. Numbers. Money. People. Computer World.
Even if you actually did do a professional-type benchmark (and I don't see how you could - you wouldn't have the resources to do it) - let's say a couple of months ago, it'd be irrelevant because Linux improves DRAMATICALLY faster than NT could ever dream off. Your benchmarks are obsoleted even before you have the time to write a paper on them.
Lastly, something that cost more is not neccessarily better. Never has, never will. It's really a pretty common sense rule, and it doesn't even apply to only the software field. It's pretty much true for everything. It does take an idiot to believe that more money == greater quality. (sorry, I don't usually insult people, but the temptation in this case was just to great. he asked for it)
I could go on why linux is better than NT (and speed is not a factor) but it's really not worth doing in this thread.
> What is this, stream of consciousness writing?
It's entirely possible some people speak English as a second language.
I've finally had it: until slashdot gets article moderation, I am not coming back.
Being that you are an expert in these things (so it seems). What would the appropriate Linux solution be? How much would that cost? You get roughly $50k to play with over Microsoft. Difference in NT -> Linux pricing, estimated at $500 (Bulk Purchage) x 96 ~= $50k. Use a better figure if you have one.
The point - High volume servers will always be expensive. Witness half million Oracle servers.
The best part about OSS or Linux or FreeBSD is it provides a CHOICE. Does whatever you are using work well for YOU? If so, then who cares what someone else uses? For the record, I've had a non-M$ home for at least 6 years now, using Linux and Macs, but sorry to say, currently FreeBSD and Macs.
Here is the ZDNet graph. According to this graph, NT4 scaled about 2.22 times and Linux scaled about 1.63 times when comparing 4-proc performance to 1-proc performance. The ideal scaling would be 4 times, or linear scaling.
cpeterso
That's the ultimate question.. Microsoft may have a performance advantage (gasp!) right now.. but we all know how quickly open source moves forward, and how quickly bugs are fixed. Even Microsoft can't beat the distributed efforts of tens of thousands of developers working in concert. No corporation on the planet can.
I don't know about that...
Linux has thosands of average hackers working in there spare time
Microsoft has hundreds of really good coders working full time
I think that the two are pretty evenly matched, at least in terms of what they *can* produce.
Dosn't solaris beat out NT?
_
"Subtle mind control? Why do all these HTML buttons say 'Submit' ?"
ReadThe ReflectionEngine, a cyberpunk style n
I hope you're not trying to say that coming up with excuses was inappropriate for the first round of tests. After all, how far did Mindcraft turn out to be off? A factor of 2?
Hmm....
The german CT-magazin (http://www.heise.de/ct) just performed a similar test but with different results. Linux beats NT in all but one category (NT server with multiple network devices is faster).
Who's right?
Sven
Shit... You got me 8|-)= Try DAS sometime; THE best beer in the world...
Maybe you didn't read the graph right. That's a 1-processor NT box that tied the 4-processor linux box.
Ouch.
I've finally had it: until slashdot gets article moderation, I am not coming back.
Although GCC is one of the most portable compilers, the RTL generation routines aren't well suited
to the register-poor x86 architecture. The main difference, however, is the code scheduler.
GCC doesn't do much P6 style optimization, where VC++ in conjuction with Vtune from intel is quite
an effective optimization tool for the x86...
Well, just get pgcc (the Pentium/P-II optimization patches) patches for egcs - I rebuilt stuff like GNOME and XFree86, and saw significant performance increases - at least 20% faster performance. CPU intensive stuff like The GIMP and XMMS run faster as well.
Yes, Kurt, you hit the nail on the head.
/.'ers who could take Linux and Apache and merge Apache into kernel space, compromise everything else, and personally release an OS/Web Server combination that could easily beat NT in Webbench. The same goes for Linux/Samba. What they would have created is an O/S with dedicated application functionality.
There are probably a hundred (a thousand?)
If Micros~1 really wants to beat Linux in general purpose operating system performance, they need to take this approach with *all* other applications. Start by integrating BackOffice, the rest of IIS, IE, Office (why restrict this brilliant strategy to server-only apps? MS should surely strive for the fastest desktop also) and their other in-house applications into the kernel. Then they will FLY!
Of course, *some* of this is actually good from an engineering perspective. Common functions that are essential to the performance of standard and widely used services -- and can be significantly improved by moving them into kernel space -- may justify this approach. Large chunks of application-specific functionality, however, will weigh down non-users of those apps and compromise stability for those who do use it.
Realisticly, what I think MS has done here is create a "benchmark special". They have picked two high-profile applications and integrated them into the kernel a little too intimately so they can claim that NT in general is faster than Linux. The actual usefulness, of the web server speed up anyway, is questionable. Do *any* sites actually serve that many static pages? And, how many of those sites can afford the instability that such approaches bring?
Sorry, Microsoft. What you have created is an NT/Web server/file server combination that is faster than Linux in those same areas. That does not make NT the faster operating system -- and it most certainly doesn't make it the better operating system. Meanwhile, you have pointed out what are now high-profile areas of minor weakness in Linux performance. Those will be fixed -- and fixed correctly. Thanks.
Geeky modern art T-shirts
http://www.kt.opensrc.org/kt19990617_23.html#9
A web-server kernel module, which is as fast as Zeus, but at half the CPU usage -- according to one variety on benchmarks.
"...Bjorn Wesen objected to the entire idea, saying that if the web server worked better as a module, it was only an indication that the OS itself was broken and should be fixed."
That sums it up: the fast & easy (IIS) way vs the Right Way.
-_Quinn
Reality Maintenance Group, Silver City Construction Co., Ltd.
From CMP's "Information Week" June 21, 1999:
"When every minute of downtime can mean millions of dollars in lost revenue, companies generally rely on applications that run on OS/390, Tandem NonStop Kernel, Digital OpenVMS, or Unix operating systems. But Windows NT is increasingly being deployed... so IT managers must find ways to increase the availability of their NT environments. To do it, they're adopting products and services that promise to provide extra protection..."
" 'Any system with lag time is unacceptable for running the application' says William Harris, NT Administrator for the Ohio Utilities. 'Money wasn't even a big deal. I's rather get quality and reliability and availability'. The organization...paid $75,000 to implement the (third party protection) system.
Translation (for those who need it): Management is telling IT they have to transition to NT. IT says, in order to be stable, we have to add third party help. Management says: "Here's a blank check."
It goes on to say that Unix, w/o third party software or service achieves "availability in the 99.9% range, as opposed to 97% for NT."
Now, what's the difference to a business between 97% and 99.9%?
IBM's NetFinity Availability Program guarantees 99.9 w/ NT. Cost: $220,000.
HP Mission Critical guarantees 99.9 with NT for a mere $300,000.
Imagine going to your boss and saying "Hey, how'd you like to save $300,000?"
JL Culp
Business Technology Consultant
Chair, LPSC
babelfish.altavista.com
cpeterso
Sure that explains how Linux lost in the multi-cpu tests but what about losing the single cpu test. What is the cause of that?
ayottesoftware.com
If I understand this correctly, you're implying that NT was modified to be app. specific to target a narrow purpose (web servers) whereas linux is more general purpose and has a wider application base to satisfy. Actually, the opposite is true.
MS has a far wider user base to afford making their OS better for just one app. at the sacrifice of other applications. Certainly not web servers, which makes very little money for them.
I find it annoying that people are coming up with all kinds of excuses instead of facing the facts. This reveals our own denial of reality, more than anything else.
L.
Although it wasn't clear from my original post, VC++ is significantly better than PGCC.
Standard GCC on x86 is a dog.
Even so, PGCC is probably not the answer either since I've heard that there are problems
compiling some kernels with some versions of PGCC (but maybe that's fixed now)...
OSXS is in it's childhood. In the fall, Darwin, OSX, and OSXS will be resynced. This will give everything a much faster and capable kernel based on mach 3.0. Other perks will be included such as posix threads. After that, it get's sketchy. OSXS will evolve to be a pakcage that runs on top of OSX. This will be similar in concept to AppleShare IP today running on top of OS8.X. OSX will be radically different from OSXS featuring a much more sophisticated interface, and the deatils of the unix underpinnings will be much more hidden from the user. Hopefully they won't take away my command line though :)
Clustering: If anyone here has had trouble running a cluster with recent versions of the aforementioned "unreliable hacks", I'd like to hear it. Any takers?
Anyhow, that only refers to failover-type clustering. I've yet to hear of a widely tested NT-based alternative to Beowulf.
Security: Quite right. Any machine out-of-the-box is quite insecure. We're referring to competantly administered systems, right?
Stability: Win2K should, based on the input of those beta testers I've spoken to, be quite stable indeed; I'll not fight you on this. As for NT4, however... well, there's quite a bit of downtime (nevermind the hassle) involved in the sp5 upgrade, now, isn't there?
Where I work, they spend more money buying sandwiches for meetings than you would pay for an NT server license.
Not everyone works out of their bedroom.
I've finally had it: until slashdot gets article moderation, I am not coming back.
1. Linux is OPEN (what Stallman means by free)
2. LINUX is stable (unlike Mac or Windows)
3. Windows (95 and NT) have MAJOR knuckleheaded design compromises. The windows UI is horribly broken (drive letters, anyone?). Linux is not there yet, but I'd rather start out tabula rasa without the frills than stick with the bloated beached stinking whale that is windows.
Given that Micros~1 cannot fix the bullshit design compromises of Dildows, it will eventually catch up with them. The windows comcept is starting to collapse under it's own weight.
Linux does NOT have those sorts of problems.
Too much baggage. Time to start over.
support gun control: take guns from cops
ahhh.. just shutup and get coding.. :)
I think that the file server portion of the test is inherently flawed in that it uses a non-native protocol under linux...SMB. SMB is native to all windows platforms. They should be testing NFS as well, then I would think that linux would fair better. This was merely a test against samba vs the built in SMB portion of the NT kernel. IIS 4.0 vs apache was much the same, test apache vs apache. To some extent Linux is also to blame. ext2 is not a mature file system, XFS should help, I would like to see BeOS open source their file system as well for our use, blatant plug ;) and as they mentioned the multi-threading of the IP stack should solve any problem there. The tests seem to exploit and weigh too heavily on these two shortcomings of linux, I'd like to see NT's wolfpack clustering compared to that of Beowolf...but I guess that's not necessary considering that NT is listed nowhere in the top 500 ;) Let's take the cue from the IP stack problem and from BeOS and multi-thread the hell out of the kernel, as that's what will make multi-processor machines fair better. That with the adoption of XFS and then we'll call for new tests against win2000 and laugh our asses off, because like win98 is to win95, it's just winNT4 in a dress with makeup, I still wouldn't do her, know what I mean?
There were some performance tweaks in SP5 along these lines, but I forget exactly what they were.
--
Ben Kosse
Remember Ed Curry!
It's been known to break some NIC drivers causing a blue screen. We had a pair of systems with this issue (2 HP LCIII systems with 1 each 3Com 905B and Intel EEPro cards where the 3Com driver would cause the bluescreen).
--
Ben Kosse
Remember Ed Curry!
That's a very interesting and valid point. They should also do a price/performance comparison on different e-mail servers, like Sendmail for Linux vs. Microsoft Exchange for Linux. Oh yeah, I forgot, Linux would win for sure, because MS Exchange ends up costing like $50 a person, where as Sendmail is free, as is the OS it runs on. Not that I'm saying NT is inferior or anything... :)
-- Terry
Everybody pretty much agreed that the initial Mindcraft tests were biased. This is why this test was held on neutral ground - with RH on hand, and with everything closely monitored.
/.ers will cry foul and hint that whoever is conducting the tests is under the pay of MS.
The problem is that no matter how neutral the test is, if it shows ANY OS performing better than Linux in ANY category,
This is a pretty childish approach. Just looking at all the posts it's funny how many different excuses are being thought of. Face it - acknowledging fair play like a good sport and fixing shortcomings will do far more to help than simply sitting and spinning conspiracy theories.
I at least hope nobody is flaming ZD like those clever guys who smothered mindcraft with witty remarks. It would do wonders to our credibility...
L.
>These studies do not address price/performance.
;)
True, but how exactly do you plan to measure performance, and what aspects of the server should this performance reflect? Perhaps units of dollars/(clients served/second)?
>These studies do not address stability.
Again, how should you measure stability? I would suggest in units of crashes/month, but then MS would cry foul
>These studies do not address security.
Cracks/week/cracker?
One thing that concerns me though, is the reaction the community will give. I think, unless some major news comes out, that we should accept these numbers at face value and implement changes in the kernel and all other appropriate software. In this respect, Linus' idea of stable kernels released more often would be a good idea; many people don't like the idea of recompiling a 2.odd kernel on a deployed server, since it would affect the stability. Accepting these values would show the world that we aren't a bunch of crybabies and are willing to put our code where our mouths are.
Depends. If you just run the operating system, it's stable. Once you start doing anything else, watch out.
I know; I've had to deal with NT for years. Hint: try running a web site using ASP and FoxPro.
Also...ever hear of SP2? Eeek.
And how about the IIS .htr hole?
And as far as costs go, you forget that most businesses are small businesses. For them, a $1000 Linux box makes a lot of sense.
--
I'd rather have CE than Java giving me chemotherapy.
sorry; exchange server... the biggest joke in software history...
I use nt/linux/solaris daily for mission critical stuff; solaris wins in stability, nt just crashes (we have a dedicated(he has those sh*t msce something or other courses, nt admin) a lot (it aint fast and it aint stable), and linux is fast and quite stable.
Any OS I can crash from userspace is crap; nt is a piece of cake; i just sit and program (does not matter which language) and it crashes at least once in a few days.
Linux is quite easy if you have an out-of-the-box-system; you can secure it enough to make it virtually crashproof.
As far as speed goes; we use linux as file/webserver and it is much faster than the nt mammots we have; this is practical experience; userstats.
I don't know were people use nt for (we tried it as intermediate db server), but it simply cannot handle the load. AIX/DB2 is the only thing we have tried that can handle the 100 mlns of requests we get/day.
If IBM is serious about f*cking M$, then they will help linux get to that point.
btw. m$-sqlserver failed bigtime on our tests.
You might not have heard, but Bill's dream
is to have the CE version of windows running
on all embedded systems in the world.
Everyone who sponsors M$ now will be part of making this happen.
Good luck to YOU when you receive chemotherapy from a win CE controlled system.
It's not about their products, it's about the worlddominationthing they want!
That's why you should think twice...
Not *EVERY company* wants world domination. Most companies just want to be the best at their thing. M$ wants to *be* everything; that's the bit that worries me. If they would say; win98 is great for games and nt/win2k is nice/cheap for groupware then it would be acceptable. But saying; we are the best at everything, is stupid.
If anyone would say such a thing, we would say; 'don't be naive', but Gates == god?
You cannot tell me that he does not want *everything*! Tv/OS/embedded/telcom etc...
That's the bad thing; and *NOT* every company wants that!!! It is plain ignorance to want something like that. When the time comes for m$ to should it out, Bill* will be gone.
ok, here's my $0.02 on Windows 2k (and, fyi, I run Linux most of the time).
:-)
(First and foremost, these are just my impressions of Win2k...not cut in stone by any means)
First, my computer is a p200 mxx, 64 megs, ~1 gig NTFS, ~2 gigs ext2. W2K found all my devices and configured then almost perfectly. The only thing it didn't get was my Voodoo 2, but I can run GL Quake in Linux
The system runs faster than NT 4 ever did. Some of you may than scoff NT 4's performance, but let me say this: I started using Linux because NT 4 was too slow. W2K (approximately) matches the speed of Linux in performing tasks (starting WP vs starting Word97). There's only one other nice change. It hasn't BSoD'd yet. Its stable and quick.
Now, for all you Linux zealots: problems w/ win2k.
Its a beta. I understand that. But it really shouldn't stop being able to look up things via DNS. Its an infrequent problem, but its annoying.
Next, it does kinda take over for you too much. I was surpirsed after a while of using W2k that my application icons in the start menu had disappeared...Windows had a cheerful message telling me that it had optimized my Start menu. I really would have prefered if I could have asked it to do that for me, but ah well. Next, I used to run NT in 1600x1200 perfectly. W2k seems to have trouble drawing at that resolution...I had to revert to 1280x1024 (fyi, its a Matrox G200 SD, 8 meg - drivers come w/ Win2k).
Conclusion. If MS can clean up the problems, Win2k will be *very* nice. Although it can't run servers up the wazoo like Linux can (than again, NT Workstations was never designed to run servers, and therefore shouldn't be tested, IMNSHO), it runs well, far better than any previous MS OS.
Note to MS: Open up the source to NT/W2K. Open Source development of NT would speed up removal of bugs, and I would think that NT would probably speed up as a result. Plus, if the good of Linux and the good of NT could be mixed together into an GPL uber-OS, I would be happy...hell, I would even pay for it...
--------------------------
...started with FreeBSD and switched to Solaris, IIRC.
Or maybe they only partially switched. .
Maybe when you were working there, you thumbed through a ZD publication or two.
What do you know? They're full of ads for Intel and NT-based products. Big Super Suprise!
--
Business. Numbers. Money. People. Computer World.
Zeus may be faster than Apache. The bottle neck was not Zeus or Apache but Linux's IP stack. I would assume that on a different IP stack, you can determine which, Zeus or Apache, is indeed faster. The result is only true for Linux.
IIS is free (as in beer)
_
"Subtle mind control? Why do all these HTML buttons say 'Submit' ?"
ReadThe ReflectionEngine, a cyberpunk style n
Keep in mind that the filesystem test is on M$'s "home turf", namely the screwy SMB network file system. (personally, I want to see how NT serving via NFS compares to Linux and Solaris) Samba was developed outside of Microsoft, by reverse engineering the wire protocol, without the benefit of any documentation for said protocol. The Samba development team should be congratulated highly for the fact that Samba performs well enough to warrant ZD throwing a tiebreaker benchmark test. This is no small feat.
Likewise, The Linux developers should also be congratulated highly for the fact that Linux is even mentioned outside of academic journals. That Linux is good enough to be measured against M$'s finest by the mainstream PC-centric press is also no small feat.
Sure we're not the *fastest*, but seriously, who was expecting we would be? The point of this round of benchmarks wasn't to prove that Linux was faster than NT, but to show MindCraft that Linux was represented unfairly in their tests. And we showed them that Linux was about twice as fast as they claimed. Rejoice, for the truth shall now be known by all.
Oh, and get back to hacking the source, we've got work to do. =)
"There are lies, damn lies and statistics" -- Mark Twain
How's my programming? Call 1-800-DEV-NULL
Currently the VC++ from microsoft produces far superior x86 code than GCC. I'm surprized that
nobody has commented on this fact yet.
Although GCC is one of the most portable compilers, the RTL generation routines aren't well suited
to the register-poor x86 architecture. The main difference, however, is the code scheduler.
GCC doesn't do much P6 style optimization, where VC++ in conjuction with Vtune from intel is quite
an effective optimization tool for the x86...
It would interesting (but unfortunatly impossible task) to find out how much of the difference is
due to simply the difference in compilers...
Just my 2cents worth...
We use NTsp3 for running MS Access queries/reports. The machines we use are factory direct Dells. We get BSODs about once every 2 weeks or more. Usually its the hp printer driver that causes the crash. Does that mean it's a faulty 3rd party driver that crashed the OS?? Maybe so, but why does the WHOLE machine have to go down after I hit the print button??
Stability would be nice.
I think it is a tempest in a teapot... Practical experience has shown ME that NT has been unreliable and slow on my nets compared to linux. Fact of the matter is, why would i run a webserver using samba? I would never do that. I run linux, and apache for my webservers. They never crash and they hand 70000+ potential users, using a .cgi script. I have NEVER....I repeat never had a crash OR complaint from my users period.
Without regard to the tests conducted at ZDNET, practical experience has shown me which webserver to use, not to mention, which one i use for more mission critical applications and services.
I suppose that those that are tired of the crashes will turn to linux. Those that dont experience it will continue dishing out the bucks, and stay awake at nights wondering if when they go into work in the morning whether or not there will be a 'blue screen ' waiting for them and then they go on.
I find it really odd that some system admins find it entirely acceptable to have a server and/or workstation crash. I would like someone to explain that one to me!
Terry Funk
I believe it is in the kernel. I just stopped and restarted my Server Service while watching the Task Manager. No processes died. None were created. It's got to be in the kernel. I wish I had my MCSE (I know, but it pays the bills quite well) training guides to verify this from one of the pretty documents.
Further proof comes from the fact that changing the configuration (bindings) of Server, Workstation, or any of the other network services requires a reboot.
Isn't that kind of shortsighted?
Linux has many advantages over NT, including increased configurability, open source, GNU utilities, etc.
A few skewed benchmarks don't mean anything. There are benchmarks out there that show Linux beating Windows NT. Benchmarks don't mean that much when there are a lot of other benefits to a platform. It's kind of bad to base your decisions on one benchmark, when you have to consider everything else that you get with the system..
Our company just switched to Linux servers, away from NT. NT was excessively difficult to remotely administer and install, and we found it to be very cumbersome, with very high overhead. Linux has so far been an absolute dream to remotely administer, and our server implementations have yet to crash or even require a reboot. Now that we have all our configuration files completed, installation is very fast and simple, leaving more time for testing any new changes.
I apologize if this sounds bad, but it is a very bad business decision to change your platform based on a single benchmark. NT is more costly, both in cost and administration. My company's experience with it has been very poor.
If you're having trouble convincing your clients of this, just point them at the Kirsch Paper. It worked for us.
-- "God, Root, what is difference?" - Pitr, "User Friendly"
Another benchmark claiming something that I find hard to believe, but it is amazing how many people in this discussion sit here and defend linux so much, here you are mainly preaching to the choir except some really insecure windows users. Don't get me wrong, a lot of the people here are insecure about linux, but the windows people are VASTLY less secure and more defensive.. Maybe These are the ones afraid that if there was a change they would be phased out for someone with more unix experience? I dunno, I tend to be a bit righteous about linux/FreeBSD for some reason so I guess I don't have much room to talk...
The fact remains, if the free operating systems are so horribly bad, why is it so many private users and corporations that use them FEEL like they are more stable and better performing? Oh well, just my righteous two cents.
XML is like violence. If it doesn't solve the problem, use more.
Have you been fixed yet?
(and no I don't condone Windows servers)
- alex
The revolution will be mocked
This article is about as sparse as possible, will they hopefully elaborate soon? I personally think nothing but good things can come of NT beating us. We do have a lot of areas to improve speed within Linux, and I'm sure we will all enjoy reaping the benefits of straightening out our shortcomings. I wonder what the results would be if they ran both benchmarks at the same time?
I guess the big question is, how many times did they have to reboot NT during the test?
Josh
Squash
Quite simply, prince toadstool. Download - recompile - reboot.
When we see where we lag, we write faster code. We post the source. We all merrily repeat the steps above. We close the gap in NT's "special areas" every day.
AND with anonymous agitators such as yourself egging us on, we shall double our efforts to close in on the evil empire. Thanks for the encouragement!
SMP shall be ours. Oh, yes...
----- if ($anyone_cares) {print "Just Another Perl Newbie"}
warn "Just Another Perl User" if $anyone_cares;
Do you always say it with your caps lock on? :P
I didn't participate in the spamming of ANYONE. I dislike the mere thought of it. Had I decided to send an email it would have been 1) to request more info or 2) to debate a legitimate problem.
If you think that *ALL* Linux advocates are spammers, children, or whatever your mind can think up, you are sorely mistaken.
Try RE-reading my post. And do it with an open mind ok? Not your "Microsoft is good and everyone else is a heretic" attitude.
That's what I read about it. My apologies. :)
But since it's Posix-compliant, does that not make it Unixen?
Fine. Why don't you design a benchmark that you think the free Unixes will do better at? I mean computer performance, not price performance. :-) One obvious thing that comes to mind is this risible Samba thing being replaced by NFS. Another is generating dynamic pages instead of static ones. But that's just the start. What else would prove interesting?
And what about running a variety of operating systems on the same hardware? What about BSD? What about Solaris for an x86?
HPFS has been helped along with the JFS from
AIX now ported to the new Warp Server. It doesn't seem to be an SMB speed demon but it scales rapidly like it does on AIX. HPFS386 was/is nasty fast with SMB file and print using busmastering cards and has even less fragmenting than the original HPFS.
I'm wondering why all these other OS's get tested but OS/2 Warp and Warp Server aren't ever tested? I mean it is multithreaded at the kernel, runs SMP quite well, and with the free EMX and XFree86OS2 you can run many recompiles Linux applications and daemons. It seems to me that it is really more likely to beat NT because of its design and you get Linux compatibility via the OSS that is EMX/EGCS/XFree86 for OS/2. It was OS/2 that outperformed NT when NT was running 4 CPUs and OS/2 was running on only 1. We should be promoting choice from all players and yes OS/2 is still being sold by IBM you just have to ask or go find it. If we knock NT down we all win even if it is knocked down a notch by that obscure OS formally built by Microsoft themselves. Wouldn't that take the cake? (I love it, QT v2.0 is already available on OS/2)
"Anyone who stands out in the middle of a road looks like roadkill to me." --Linus
NT's kernel was designed for symmetric multiprocessing from day 1. Every part of the kernel was written with thought given to whether or not it needed to be locked against another processor.
Linux, on the other hand, was originally written for single proc. As I understand it, they only recently started supporting SMP -- and then by having large granularity locks that keep multiple processors out of huge sections of the code at a time. (The article talked about Linux having a single lock around the whole TCP/IP stack!) To fix this, you basically have to go over every line of the code and only lock the things that need to be.
The interesting thing to me, is whether the Linux development model will support this well. Writing SMP code is much harder than single proc code. All those race conditions, deadlocks, and missed data contentions to worry about. People really have to understand what they're doing to get it right. Already there's complaints about the 2.2 kernels not being as stable as the earlier single big lock kernels.
Of course lock granularity doesn't explain the whole picture. NT still trounced Linux pretty badly in even the single proc case. There, I suspect it's just a matter of Microsoft having a greater number of highly qualified people working on the system than Linux does. Not that Linux doesn't have any highly qualified people, but rather that MS can get more of them. Paying people for their labor actually seems to work sometimes.
How many different servers can you put onto one machine with NT? With Linux? What kind of performance do you get when you have a mail server, DNS server, Web server, etc. all on one machine on NT versus Linux?
These are all things to consider before dismissing Linux because of one benchmark.
---
This sig has been temporarily disconnected or is no longer in service
I don't know what reality you live in. The reality I live in, which includes several million page views a day, would not follow the pattern you're describing at all.
I'm NOT by any means an NT bigot... where I work, we use FreeBSD, but just because your web site sees very little traffic, don't assume that you are the "standard". Hobbyists, in the grand scheme of things, are less important to "document and benchmark", than business/e-commerce sites are.
For the hobbyist web-server who cares if the page is delayed by 0.1 seconds per page or if the server takes a shit on 100 req/second. It won't happen for the average person. Benchmarking's purpose in life is "Here's us beating the living bejeezus out of the machine .. at what point does it break"
Essentially...
Redhat had some of its best people at the test lab when they were being run. The tests were closely monitored by MS, MC, and RH.
If the linux community keeps crying foul under these conditions, it will only hurt our own credibility.
It's time to stop whining and start fixing.
L
Computer system design is ultimately a vast series of tradeoffs. M$ has, in many cases in the NT kernel, traded stability for speed. That is why NT's uptime numbers are so comical. NTFS is probably just plain better than ext2fs, also. SGI's XFS contribution may alleviate this, however. I estimate that it will be 1 year at the most before Linux wins decisively in all relevant categories.
If Microsoft can stack the deck, we can too.
NT IIS vs NT Apache vs Linux Apache vs Solaris Apache
Hardware: One PIII 550 w/ 768 megs of RAM (Fill in same level hardware to make functional system)
8 tests:
I think that would give us the results we wanted for optimal publicity, but if I'm wrong . . . give a better set.
"Well, if NT won't boot then it forfiets that test"
-- The act of censorship is always worse than whatever is being censored. Always.
It was my understanding that the server service was the NT equivalent of samba ... NT certainly doesn't serve files without it.
support gun control: take guns from cops
Benchmarks are good for one thing: Benchmarks.
Benchmarks are fine and great and all, but in all my personal experience changing servers from NT to Linux gave everyone a performence increase... I know this is mearly anecdotal evidence at best, but that's what has worked for me.
[Silly Analogy]
As for the samba tests.. it's something like this: Microsoft makes up a game. Microsoft doesn't tell you how to play the game. You try to learn the game... Microsoft beats you by a little.
[/Silly Analogy]
Of course, this test doesn't show reliability though.. how long could they each handle those loads? Just the (what hour?) time it took to run the test or 24x7 for 6 months....
Anyway, to incorparate everyone other post we'll see: Well, we'll get better.
not meaning to sound immature, but its true
linux isn't rocket science... NT is simpler, but still...
for just a second...
"What fails to kill me makes me only stronger."
-F. Nietzsche
Thank you. Now quit crying and start coding.
That is all.
Should hold one of these benchmark test every six
months and invite more contenders. I would like to
see FreeBSD, Solaris, Irix, and even say BeOS
in these. It could be done during a convention weekend and marketing could have a field day.
Well, unlike what you say you know. I use linux, and I know how to use command lines, I'm not a mouse crazy idiot. But I do know where you draw the line. I'd rather use the mouse to do tasks which would be faster if i use the mouse.
All Microsoft apps are heavily bloated? I don't think so, for what they do, they are rather slim. Try making an appliation do something, then try making an application that does something, and makes it easy for joe bloggs to use.
And I didn't compare Linux users to MS Programmers, I didn't even mention the word programmer in the sentence.
Stereotyping Windows as a file eating, memory hungry OS that is no good for nothing is jsut stupid. There are millions of copies of Windows out there in offices and homes, which work perfectly well for those people.
And so working for a company polutes you cause you get paid?
I guess work isn't that great now...I guess Linux shouldn't be working for transmeta, maybe he should do what he does independtly cause the quality of his work would be better - since he wouldn't be getting paid.
Forgive me, but I'm confused, perhaps it's caue I don't work with Linux programmers, but what programmers do you know that aren't proud of their programming work, and wish to makeit better and cooler, regardless of whether they get paid or not. Microsoft certainly makes a better working enviroment for their employees, who generally get to do what they want to a point.
Think about raster or alan or some other major Linux programmer who gets hired by RedHat, does that all of a sudden make them not willing to write decent software?
I think this feature explains, at least in part, NT's superiority in multiple-CPU raw service.
A side note to flamers: please, PLEASE don't treat these results as suspect or corrupt. I don't think they are. Don't think of them as a defeat, think of them, like ZD said, as a roadmap to show where Linux needs improvement.
Actually, it may be difficult for some of us to imagine, or understand, that anyone speaks a language other than English. Period.
-rozzin.
I don't know that much about NT, but someone
commented somewhere (maybe on Slashdot, maybe
on Usenet) that there are web servers geared for
speed (Boa?). Shouldn't one of them have been
used instead? After all, Apache's not built for
speed. What was the NT server built for?
In theory, theory and practice are the same; in practice they're different. (Yogi Berra & A. Einstein)
I used to run these benchmarks, and worked with the people who wrote them. They were designed to work best on Intel hardware with NT and IIS. It was intentional, and to use a ZD benchmark in this type of comparison is laughable.
------------------------------
Kinda reminds me of the strip. Sometimes you
get beat off the line but you know you have the
power.
I've been in places when I was two car lengths
behind at 1/8 mile but I still had a big smile
on my face 'cause I knew I was gonna' catch him
and blow by.
I think that's where we are now. NT has come off
the line and is maxed. We came off a little wonkey
and are just now getting traction.
It' just an analogy but I think a good one.
CC
"Pray arm me further by your reply" Winston Churchill
OK, for the most part, I'm inclined to agree. But I'm confused about something:
Poor drivers aside -- both OSes can have trouble here, so I see this as a poor defense for NT. (Why are NT's drivers poor? Doesn't NT's interface, etc. make development easier?) -- if NT can be made stable via better administration, why is it that so many people have stable Linux boxes, but at the same time, can't get NT to run worth a damn? Isn't NT easier to administer? Can't a dummy do it now? (some of those books really rub me wrong...) This seems paradoxical to me. I see many people lately referring to arguments against NT's stability as "cheap shots", etc.; yet, experience (my own, and others) doesn't seem to correlate this. On the other hand, I've found Linux to be quite stable in comparison, while I typically find NT to be "cross your fingers" propostion. Sometimes it's OK, but not nearly as often as Linux. Yes, even when using hardware on the HCL, with the latest service packs, yadda, yadda -- for good admins those things are standard procedure on any OS. Seems to me recent arguments that NT's lack of stability in some situations is not NT/MS's fault are contradictory to claims of it's being so much easier to use. It just sounds like a lame excuse to glaze over a real problem.
"I don't think I ain't" -Thompson's Corollary to Descartes
NT has put many services in kernelspace and has largely bypassed their HAL in favor of multimedia performance - especially video.
NT uses a multithreaded process model for IIS and SMB file-services that results in higher throughput but less stability. A single thread of the main process may die without completely destabilizing the server but if the main process dies then all child threads die.
Linux divorces the graphical user interface from the kernel thus ensuring stability (framebuffers are available for video enhancement though) and implements most services as userspace daemons.
Linux uses the forked process model to provide services to multiple users. This modem achieves stability in that if one process dies, the others continue as if nothing had happened. Both Apache and SAMBA operate in this way I believe.
NT has chosen performance over stability.
I believe that with kernel enhancements and profiling, any bottlenecks in the networking system can be eradicated causing Linux to perform much faster and possibly even beat NT in tests such as these.
Codifex Maximus ~ In search of... a shorter sig.
I don't know that Samba and Apache selected the forking model by choice. Not to long ago threads really didn't work under linux, so there was no choice. I had a similar experience while working on a video game called Golgotha. We *really* had to have threads, but they didn't exists or didn't work properly.
:)
The kernel has had fundamental support for them for a longer period of time, but a reliable thread safe libc (and libX11) did not surface until recently (the last year or so). Even now the distributions are struggling to get everything converted over to the new glibc.
Even though linux now has good thread support, gdb has trouble with them (last time I checked, which was a while back). Also, apache and samba are not linux-only products. The safest and most portable approach to take back then was to use a forking model.
Other than taking up an too much memory and thus causing swapping, I don't see why a threaded model would be any faster than forking model if pooling is used. Pooling keeps a fixed number of processes running all the time and dispatches request to them. When they finish they sleep and wait for the next request. This way you have the safety of a separate process space and you avoid the forking (as in you "forking piece of sheet") over head. You let each pooled process have a fixed lifetime in order to clean up any leaked memory. Some systems have libc leaks that can't be avoided, so this is important to long-term stability.
You might argue that a thread context switch is less expensive than a process context switch. Under NT (and 95), all threads are scheduled without regard to what process they belong to, so at most could skip some page table changes when going from one thread to the next. I doubt that is the case, because threads jump from ring3 to ring0 and back while executing system functions (ensuring they have different page tables). Also threads have a separate segment mapping for fs. A 3->0->3 context switch is accomplished by an interrupt in NT. One way to speed things up in linux might be to have a way to pool system request up in ring3 and then dispatch them all at once in ring0. NT did this with their GDI code. Doing this for all system calls wouldn't require a huge change in the kernel, but it would make user code harder to write as system calls have to be parallelized. GDI doesn't require any changes to user level code, because there are no return codes to worry about for graphic drawling operations. Xlib has the ability to que up commands as well. But that is not going to speed up apahche.
Another possible advantage to using a thread model is faster IPC. With something like Apache, there shouldn't be very much IPC going on except to dispatch a new request and possibly lock protect common files.
I have not looked at a single line of apache code, so I can't say for sure, but there seem to be a number of httpds running all the time, which would signify that it is using pooling. So why is it slower than the NT counterpart?
-- Virtual Windows Project
Microsoft posted a comparison on their site that showed how Linux is more expensive than NT when you include hardware into the formula. The formula MS used included OS, http server and hardware costs. But they forgot a few things...
If a company is ready to spend $60,000 on a web server/Internet gateway, they would also like to have firewall and traffic shapping capabilites. These features are much more expensive than the OS and could cost up to $60,000 alone.
Put that in the formula and see who ends up winning the games of value.
ayottesoftware.com
It appears that under these conditions NT has outperformed Linux. Now let's repeat the tests every 3 months for about a year. On the same equipment. Oh yeah, leave the machines running.
anybody know how bsd would fare in the web tests?
But here is our chance to show the world [and MS] why Linux and other OSS projects are such a good idea. By quickly implimenting fixes to the problems brought to light by these test, we can prove how much better OSS is.
Proposal: Annual or semi-annual benchmarking of NT [or the current MS server platform] and Linux [and any other OS's that want to compete I suppose]. By doing similar tests regularly, we can show how efficient OSS can be at fixing current shortcomings [as if 24hr bugfixes aren't enough].
Just a thought. ;)
BTW: Sorry for the overuse of the "OSS" buzzword
I'd help with implimenting fixes myself, but I'm not exactly an expert coder [I don't think "Hello World" will help Linux beat NT]
Ender
If at first you DO succeed, try not to look astonished!
Nothing to see here
---
This sig has been temporarily disconnected or is no longer in service
Give me your bosses email address, I've got this guaranteed way of making money.
Deleted
But I don't think that this means that speed for web serving should be any more important. Getting back at Microsoft is not a reason to improve Linux in my book. There is are many other fronts that Linux heading toward like the desktop, embedded devices, and hand helds. I can imagine that if Linux is tweeked for web serving more than normal that some test will find Linux useless for embedded devices or something else that is important.
Microsoft right now sees Linux as direct competition as a server. It will be nice to see Linux compete back but don't expect NT to stand still. There are other servers also. How does Linux compare to Mac OS X?
And no more excuses. Linux is not the fastest. Deal with it.
For now.
--
First, I respect both NT Server and Linux(Red hat), but those are not the only servers available. FreeBSD is in my opinion a very good operating system and is worth a try to see if it could handle the stress. ;-)
A note to all linux users: Don't let NT win, you have the ability to correct (not patch) the problems with the current kernel and blow NT away next year. Re-writing code never hurts, and it usually works better too
It really doesn't matter how you say it, as long as you use it ;)
--------------------------------------------
Please give your mod points to others, Im at the cap. They will appreciate it more
It's possible your personal experience with stability was on the far end of the bell curve - just because linux is more stable than windows on the average doesn't mean every linux machine is more stable than windows; the opposite is also possible.
As for poorly written drivers, it doesn't matter that the cause is that windows has to support more hardware - that's not an excuse for releasing unreliable software.
About linux support, I find it very surprising that you've gotten bad support. Normally people are very eager to help - the local user group here will install and fix problems for you in person if you bring your PC. Once you get the hang of asking questions on mailing lists or usenet you should have no problems.
About Win32 API vs. KDE/Gnome: KDE and Gnome aren't APIs. They are GUIs.
L.
You are a fucking idiot. What are you one of
those wannabe hackers who switched to linux because you wanted to look smart, teardrop windows machines, talk like a "31337 hax0r", and say "i'm cool because i use linux" (even though i know nothing about linux and the open source movement)
Get a life you moron, linux does not need or want dipshit users like you.
A gentleman who goes by the name of "CoolAss" tells us to grow up!
--
How about it ZDNet? Let's have the gory details. Specifically what tuning was done to each of the operating systems? How was the software procured? Who installed it? How was the hardware procured? I want to see it all.
"the open-source community ...claiming that
objected to Mindcraft's use of the Apache
Web server
using the fastest open-source Web server,
Zeus, would improve results."
Maybe ZD has the source to Zeus, but not the rest of us.
I do find it interesting that they determined that Zeus' performance "peaked almost exactly where Apache's did."
None of this is very good for the proposal I'm writing for my company to replace Apache with Zeus.
-fb Everything not expressly forbidden is now mandatory.
Yeah i know it's freeBSD, but they are both free unix
Even more to the point: They are both unices! User friendlyness aside, I can find 1500 things that annoy me day in and day out about windows. Most of them have to do with the interface "taking over" and doing stupid things it has no right doing. Like autoloading CD-ROMs. That's just downright annoying, especially when the company that made the CD only puts up an install option.
Lowmag.net
That's great! Good for you.
"In the other corner, two cardboard boxes; one labeled 'Windows NT Server,' the other 'Microsoft IIS'..."
This all inspired by:
(Yeah yeah, apples to apples...)
-m
Part of it is the fact that smb is built into the kernel. I've been running nt vs. linux tests at work and what happens is that samba forks a proc. for _each_ client that is connected. Imagine running a performance test with a large number of clients, and you end up context-switching as requests come in from different machines, and context switches are expensive operations!
Deepak Saxena
Project Director, Linux Demo Day '99
Deepak Saxena
"Computers are useless, they can only give you answers" - Picasso
> "Here's us beating the living bejeezus out of .. at what point does it break"
:)
> the machine
so linux runs out of bejeezus before NT? well, somebody put more bejeezus into linux then, chop chop!
geez, this is obvious!
dave
This doesn't matter much. I'm not even sure this is a good configuration for most places, 2 duals could very easily be better than one monster quad. Or perhaps fewer nics... If they want to say they beat us here then they can have it, I think linux performs much closer if not better than NT in more common environments.
I think what most people are forgetting is the important part of Linux is not its speed. Its definitely not how fast it goes in any benchmarks.
Its the fact that we have a system that we can play with till the bitter end. We can see how any piece of it works, from the low-level memory management to the desktop environment.
I know the most important thing for me has been the incredible community that I feel apart of, that have spurred communities like slashdot and projects like KDE and GNOME.
Everyday for the past 5 years, I've seen advance upon advance from Kernel 1.0.9 as more and more features have been added. And the power that open-source has provided.
And after I started learning how to program, the thrill of putting open source to the test, and being able to take the source code to an application, sit down and find the bug that was bothering me.
These are the things that make Linux the WORLD'S BEST OPERATING SYSTEM. So what if NT can spit out static web pages 4 times faster. I know my linux box can handle the dynamic stuff much more reliably and with a lot less horsepower, and if it can't then I'm looking forward to the next year or so as I watch the wizards take care of all the problems.
And as I sit there and watch, I can only hope that one day I'll be able to do my part, and that is the magic that makes open source so wonderful. Everyone that touches it WANTS to do their part in this amazing journey. You think anyone gets that from running Windows? I think not.
Lee Nevo
UNIX Admin
Kudos to everyone for the BEST operating system ever!
Posted by John Hayward-Warburton:
But! Until the bottleneck (which is being sorted) is reached, the results are identical.
So, for non-thumping applications today, standard Linux is just as good as NT and is FREE (and open). And, soon, when the bottleneck is fixed, it will be equal with the leaders.
In any case, was there a benchmark for "hacker satisfaction"?
Far too many excuses. Far too many "This is unrealistic". Far too many "The benchmark is biased."
SHIT!!!
Why can't people except losing. Linux is not fastest. Linux is not fastest. No way. No how.
I rarely cuss but enough is enough. What is this? Some kind of game where you distort the truth until it fits your expectations?
I really don't want to hear anyone call out FUD again.
--
Nevertheless, it would be naive to pretend that these benchmarks will not be used by M$ to condemn Linux as slower than NT across the board. They will do it, and it will be a bald-faced lie.
DFL
Never send a human to do a machine's job.
Well, unlike what you say you know. I use linux, and I know how to use command lines, I'm not a mouse crazy idiot. But I do know where you draw the line. I'd rather use the mouse to do tasks which would be faster if i use the mouse. :) :)
Granted, if it's possible to do it in under 200 megs (my Win partition takes about 210ish of the 450 I allocated it, custom install). I don't see this in Window's future.
All Microsoft apps are heavily bloated? I don't think so, for what they do, they are rather slim. Try making an appliation do something, then try making an application that does something, and makes it easy for joe bloggs to use.
Hmm.. so if I want to make a compiler for instance, I should make it so simplified that some shmuch who knows jack and is too fscking stupid to know it anyway can belch at the computer and churn out mind-blowing software? Come on. If someone can't do something (my example was admittedly very exaggerated) they have no business doing it. If they are learning it, have an interest, and most importantly an ability, then they are more than welcome. And let me be the last one to flame them. Unless what they do sucks...
And I didn't compare Linux users to MS Programmers, I didn't even mention the word programmer in the sentence.
It was implied heavily. If I misuderstood completely then disregard it.
Stereotyping Windows as a file eating, memory hungry OS that is no good for nothing is jsut stupid. There are millions of copies of Windows out there in offices and homes, which work perfectly well for those people.
Oh, you mean the AOLosers, the ones that think Hotmail is the best thing in the world? Those people shouldn't play with the settings on their stereo, much less operate a computer. About Win, it *IS* bloated. I have an executeable compressor here. No special drivers need to run for it to work, it just compresses it in a runnable form. It knocked off several megs from my larger Win progs. When I start Win it takes ALL of my RAM for itself. All. It makes my progs use SWAP. *growl*
And it's not good for NOTHING, it's just that it aint good for MUCH.
And so working for a company polutes you cause you get paid?
Nope, by no means. Working for Microsoft pollutes you. To varying degrees, I should have said.
M$ and their ilk (other companies with similar policies) drag down those who would be giants in other aspects than the dollar sign. I am somewhat fearful, truth be known, to develop my better ideas due to Microsoft and others like it. I fear my ideas would get stolen right out from under me and there would be a total of jack SHIT I could do to retaliate.
And I highly doubt I'm the only one who feels this way.
It's simple to avoid AC's: just browse at +1 or +2
I'm sure every boxer trains as hard as possible when going up against a tough contender. It doesn't make them "afraid" of their opponent.
-----BLAH BLAH BLAH BLAH BLAH BLAH BLAH-----
Blah.
rooooar
Out of the box, Linux still beats NT.
What about in the box?
Sorry, I couldn't resist.
+--
Given infinite time, 100 monkeys could type out the complete works of Shakespeare.
+-- (Score:-1, Moderator on Power Trip)
Microsoft is going to have some real grist to put in the FUD mill for a change. In the short term, we can expect this cost Linux a significant amount of mindshare. Over the next few months I think we're going to see fewer "Business adopts Linux" stories. Never mind the fact that most networks won't see this kind of throughput and that stability wasn't addressed, in the minds of the uninformed Linux is going to be a low performing second best for awhile. I think the real danger is that the community will feel pressured to "fix" the problems before the fixes are stable. That would not be wise as we can still tout stability as a talking point.
In a way, I suppose, this benchmarking was a gift. Linux usually gets problems fixed well and quickly when they are identified. It's probably better in the short run for Linux to take it's lumps and get things fixed even if it does force us to lose a few converts and rethink our opinions of it's strengths.
At least there is a bright side. I don't think these tests did anything to vindicate Mindcraft. Linux still lost but only half as badly. There are no good reasons for anybody's opinion of Mindcraft to go up.
Anyone know where/if more info is available on the tests? Like how many clients, what type of clients, what exact hardware, etc, etc...also, what about numbers like CPU utilization and interrupts/sec?
Deepak Saxena
Project Director, Linux Demo Day '99
Deepak Saxena
"Computers are useless, they can only give you answers" - Picasso
Whether it was low-end or not is up to debate, since Ziff Davis decided not to divulge the specs on their test machines. All we know is that the 'low-end' machines were single-processor with 256MB RAM. They're talking about 'low-end' in the giant-corporate-server scope most likely, which is different from what most normal users would consider 'low-end'.
Unfortunately, this story isn't on the web; I guess it's only in the hard copy magazine.
Can anyone in Germany provide some more details?
--
the thing is that the big (read: important) companies who want what they perceive to be the best solution will throw whatever money at the problem they have until it goes away.
even a savings of 1/2 million dollars on software licensing would be something like 7-10 competent sysadmins (i don't really know the going rate for sysadmins these days...). for a company that might have anywhere up to 100 computing facilities that need to be managed... staff costs rule the day.
so, anyway, my point is that cost really doesn't matter to the larger folks in town. important for the small companies to be sure, though.
well ...
;-)
...
... the strength of linux is that
if one is honest, (s)he would have already
guessed such a result.
linux wasn't "designd" for this kind of hardware.
BUT we all know that it will be in future
SMP & huge amount of RAM are not handled well.
there is no upwards-scalability in the kernel
e.g. you need other algorithms for SMP (>2 proc)
and RAM >1GB not just parameter changes
this is also clearly shown in the test and my
personal experience is the same...
however
development goes on fast. if there were more
2GB, 4x PII 500 machines out there linux
would have won. now it has to adapt to the
"bigger" hardware.
CU,
Armin
Just set your score threshold to one, and all of the ACs will magically disappear
Why? The point was performance, not pricing, and not even reliability (another Linux selling point).
The thing is, nobody cares. With the blistering speed that OSS develops, anybody with at least the mental capacity of a newt knows that OSS will overtake all others. It's only a matter of time. Linux development has only begun to take off. What happens when SGI starts merging Irix with Linux? How far will Windows be in 2 years? How far will linux be in 2 years? What about 10 years? What would a benchmark look like between linux and NT 4 years ago? What will it look like 4 years from now? Can you still not see the big picture? 4 years ago nobody would have believed Microsoft would be funding benchmarks against a piece of freeware. Now they are. Why? What are they afraid of?
--- A Jesus Fish eating a Darwin Fish only proves Darwin's point.
I code on NT every day and work on Linux every night. The NT box has a faster processor -- PII-350 vs. PPro-200 -- but otherwise the boxes are fairly comparable -- 64MB RAM, etc. It's possible the NT box has a 100MHz bus.
Linux is vastly superior in comparison to NT in my environments. There is no comparison, on any front that I can think of:
In my world -- comparing the two as workstations -- there is simply no comparison between the two when it comes to speed, and I don't care about anyone's benchmarks if they claim otherwise.
And shall we talk about blue screens? And memory leaks? Meanwhile Linux keeps going and going and going, staying out of my way and letting me work. Again: no competition. Linux wins going away.
This has been my experience with NT. It has not been pleasant. It is not "good enough." It is pathetic. So I don't care how many benchmarks it "wins" (especially when those benches run counter to my own observations in the real world). Linux is a breath of fresh air. I'm sure that almost any unix would be too.
Bill Gates will never get another dime out of me so long as I have any say in the matter. His products are an atrocity.
DFL
Never send a human to do a machine's job.
You might be making reference to responsiveness rather than throughput. (i.e. Bandwidth vs. latency.)
These benchmarks measure throughput.
First off... nick names are just that... nick names. You can attack me personally, but you can't avoid the facts.
Second, I am not talking about the Mindcraft tests. I am talking about the PC-Week tests, which all who were involved (including Linux experts from Red Hat) said were fair.
You are proving my point with every reply...
Pricing does not matter?? since when has a large company _not_ looked at the bottom line. For example, given 5 grand, set up Linux and NT boxes for intranet web serving and e-mail. Well, I can get a nice dual proc box w/ gobs of RAM and free software, or i can get a Celeron and pay the other $4000 for unlimited user NT licenses. Do some benchmarking on that. As for the scale of these (Mindcraft/ZDNet) benchmarks, i sure as hell wouldn't put thirty grand into an intel box (NT or Linux), I'd spring for an UltraSPARC or an AIX box. Let's bench NT against that. Then compare the reduncancy and reliability. This is a huge deficency of the popular computing press, I'd like to see them bench (all the solutions) based on bottom line price, not the same hardware. What's the fastest solution for $500 (my guess, one contender: free unices) for $1000 or $2000 then for $10,000 and finally for 30-50 grand. My guess is that then, Unices would hold the upper hand every time.
Linux has one huge advantage that ZDNet and friends are completely ignoring, the ability to put all those old computers in the corner to good use. I'd like to see them do a story (or even mention) that Beowolf cluster at one of the national labs that was assembled completely for free, with old donated computers. Last i heard, well over 100 486es and a handful of Pentiums. This is what Redmond and everyone else in the industry fear: the ability to get a job done without spending big bucks. As if money makes the world go round.
Reid G. Ormseth, Esq.
So far, I have some info on how Apache was configured. Kernel is next.
Previous comment should be (Score:-1 flamebait)
I work at a major computer manufacturer and my job is testing the stability of our enterprise storage solutions under different configurations. RH 6.0 Linux, NT 4 & 2000, NW 4.2 & 5, Solaris, etc.
I have done testing which shows that NT is less stable under heavy load (Heavy I/O for extended period, Linux has seen load averages well above 100 for extended periods without problems, NT quite often BSODs when the tests last at these levels last for extended periods) than Linux, even when using some 'beta' Linux drivers for our controllers.
This is not FUD, it is the truth.
It is ignorance, like what you are spreading, that is keeping Microsoft's pockets lined.
Some evidence?
Server: runs custom accounting package with 10 users + mail + print spooler + SAMBA + DNS + light HTTP + routing + fax server + monitoring. No special hardware.
Uptime: except for the one occasion where the drives shut down because the airco failed, it's been in continuous use for almost 2 years. No unscheduled downtime at all.
It was a replacement for an SCO Unix box. It never went down either.
These experiences are typical. So when I sell a Linux box to a customer, I guarantee that it will stay up indefinitely (given enough money for adequate cooling, UPS's and so on) . I know that it doesn't cost me anything to guarantee this.
BTW, 99.9% uptime is a joke. Continuous uptime, with no unscheduled maintenance is the target.
Do you guys publish any of your results? Do you know of any places where published reliability tests? Unless anyone publishes this, it's just a bunch of meaningless words.
Reid G. Ormseth, Esq.
I've never used Visual Basic, but I'd have to assume that Tk (with the scripting language of your choice) would be just as easy to use to "quickly develop a GUI application". And for a larger project, why wouldn't someone choose Qt over MFC?
.sig--if I want to say something, I'll type it in myself.
I don't have a
Can you explain how you managed this?
(Botched kernel modules? Accidently writing to /dev/kmem as root? . . . . )
"You've crossed my Line of Death!" "What? No! Where is it?" "Here in the fine print...."
On the main page of their test, zdnet state: 'despite significant tuning improvements made on the Linux side, Windows NT 4.0 still beat Linux'..
They didn't, however mention the fact that they formatted the fileserving partitions into 4 separate partitions to improve WinNT's performance on the front page, did they?
Although I can accept that Windows NT might possibly be able to beat Linux, the wording of that reveiw doesn't make me particularly confident it was 100% un-biased.
On a completely off-topic note: while i was editing my preferences the number of comments on this story more than doubled, in about 5 minutes. wow.
Well, it least this is going to teach some modesty to the GPL or die crowed. Sure Microsoft sucks but it seems it doesn't sucks that much when it doesn't make a BSOD (Blue Screen Of Death). Maybe it's time for the hackers here to try some more innovative designs instead of sticking to the same stuff that has been around in Unix for decades. For example why isn't Apache heavily multithreaded like IIS is ? This is much more effective with multiple CPUs.
Maybe it's the whole Unix thing that we should throw away and start from scratch a new OS. BeOS proved that starting from scratch brings good things, and speed is one of them. Unix is 30 yo... nothing is eternal, not even Linux. You can't upgrade the same technology forever.
This benchmark is about as valid as a for loop that stays entirely within the CPU cache. Not likely.
(That said, we need to fix it just for bragging rights. But anybody who bases a deployment decision on this kind of nonsense deserves what they get.)
Rob
This is a weird, amped-up benchmark most closely approximating a really small but insanely trafficky intranet. ;)
I certainly do not routinely see NT boxes performing in such a manner in the real world- and I think it's a very fair question whether even these crazy 4-way 4-ethernet-card monsters would stand up to real world conditions acceptably.
I understand one issue is latency- in other words, if it is faster for NT to serve 200 pages to one place and have another request sitting there for 20 seconds, it does it unhesitatingly to get the numbers measuring higher. Apache apparently is much more willing to pay attention to that one request sitting around getting old, and to balance out the load so that nobody gets too lagged. Of course, this is not being tested for.
This has nothing to do with MS having better people: it is almost entirely due to tradeoffs being made entirely in favor of benchmarks just to get to a place where they can produce numbers like this and have people saying, "I suspect it's just a matter of Microsoft having a greater number of highly qualified people working on the system". Never forget that the benchmarks are by their very nature an exceedingly narrow view of what the job really is. As such, the numbers become meaningless- not only meaningless in the sense of 'I don't care, I'm sick of rebooting the thing', but meaningless in the sense of producing realworld results that measure up to what the benches suggest. It strongly appears that NT servers are capable of flurries of extreme activity, but also lag pockets and serious unreliability issues- in other words, even if the machine has not crashed, your chances of getting guaranteed good response are not that great- the NT server is busy running around serving something it has cached to people in line after you, because doing that increases its benchmarks drastically. This consoles you not
OK so NT can serve twice as many GIF files as
Apache/Linux per second, so what? What
about *dynamic* database driven web pages??
Again I complain that this is not a real-world
test of server capability -- if you think NT is
so great than how come Hotmail is still being
served from Sun machines??? Because NT can't
handle it -- I rest my case.
...but still yes, this test does show where the
Linux kernel and Apache need work, I suppose.
The real thing that makes NT faster is that the entire benchmark (Hardware, type of testing being done etc.) was chosen to match NT's strengths and Linux weaknesses. On purpose.
NT is not faster than Linux. It is faster on this particular exotic HW configuration running fairly ridiculous benchmarks. But note that in all the PC Week headlines these details are dropped, and "NT is faster".
Nor is Mindcraft aquitted of all crimes, no matter what Bruce will try to claim. They were paid by MS, rigged the test in their favor, and presented it as if it were an independent test. That is lying, once you've done that as a testlab you lose all credibility.
Microsoft will just keep inventing benchmarks that happen to make NT look better. Nothing can be done about it, other than observing that those benchmarks will become less realistic every month.
and see what happens! The single processor box in the test had 256MB!
What *I* would like to know is... How do both OS's perform at their minimum configurations, how do they perform over a range of configurations, did either operating system crash during the course of the tests? What is the overall price/performance ratio?
Let's have the whole story!
Codifex Maximus ~ In search of... a shorter sig.
It works for me fine.. SuSe, RH, W2k, NT they all setup fine on my system.
> That is, Linux is LESS STABLE than microkernel architectures because a bad driver or module can't bring down the system. Linux performs better than microkernel systems.
Weeeelllllllll.... depends on the microkernel.
NT clains to be a microkernel, and from what I've read of the design docs, it kinda sorta is. But it loses on the device driver front, because a bad driver will bring down NT every time.
I've finally had it: until slashdot gets article moderation, I am not coming back.
Look like a much more fair test this time, and it has provided us with better information.
Of course, I'd have preferred to see Linux win now, but it doesn't make me feel bad that it didn't. It will also help with NT tuning information, for those of us with both.
I'd have like to see a little more qualification on what the results mean. I believe the tests only measure throughput for example, and not responsiveness.
All information is useful. I believe Linux will catch up, and far faste than NT would have had it been the slower one.
These benchmarks will always leapfrog each-other. Witness the TPCC numbers race. We all benefit in the end.
We're moderating people down now for being redundant? What's next? Moderating down because of the italics tag?
Go ahead, moderate me down.
+--
Given infinite time, 100 monkeys could type out the complete works of Shakespeare.
+-- (Score:-1, Moderator on Power Trip)
Benchmarks mean scratch, let's do some "real world" tests, shall we?
Let's make a "security" test, see which webserver can be exploited to hack your system.
But that wouldn't be fair, is it? afterall, nobody cares to hack you in the real world,
while if you server 2x static web pages, you're the king.
And uptime?
Doesn't mean shit.
Afterall, in the real world, who cares about making it servers run all the time,
they just want them to run for small tests, like this one.
And features?
Forget it, who needs PHP? or mod_perl?
And kernel features such as ip masquarading?
appearantly, nobody needs these anyway...
not in ther real world.
And the price?
doesn't count in the real world, because nobody pays the bill, per user.
Let's leave prices for our boss to worry about...
And they were so gracious letting us have NT clients.
What about unix clients?
let's see how can NT scale with NFS!
Let show them we are good with benchmark,
since it is exactly like the real world.
I, however, don't give a damn about the results.
So Bill's penis is larger than linus'? get over it.
we will show them we are better - in the real world.
---
The day Microsoft makes something that doesn't suck,
---
I'm going to live forever, or die in the attempt.
Nope. I'm part of a development group and we do not make any of our findings public.
Too bad though. There is some really useful (IMO) information that we come up with, and I have been wanting to use the resources available to me to assist in Linux development. I am still exploring this.
In these lab environments you really get to know what makes a particular setup fsck up...
1. They claim that best hardware for Linux was chosen.
2. This configuration isn't 'exotic'. AFAIK It is _typical_ for big network server or web server ( not the one you use in your garage )
3. NT IS faster than Linux, even on single-processor machines -> TCP/IP stack problem isn't the only one.
4. Mindcraft tests were biased.
It's not a linear scale.
And Linux is based on Unix, which has hasn't exactly cost nothing to develop.
>Sure, there are you guys out there who don't want
>things to be simple, you'd rather excercise your
>brains doing "hard" things like mounting
>NFS/SMB dirves by typing rather than doing it in
>a few clicks.
It's quite obvious you've never supported a large network running MS networking services.
I get to investigate when pointing and clicking don't work. I haven't seen any uglier protocols. If you're a MS fan, this isn't the technology you want to refer to to make your point.
If I setup my server using all the optimizations and tricks that the MS team used to get their winning numbers, would MS still provide me support for that machine? I doubt it.
Can I even find out the tweaks they used?
A better comparison would have been Linux and NT installs fresh out of the box with no tweaking. I don't have the time to keep track of such tweaks to my systems. I try to keep them as close to fresh installs as possible so that I can maintain them easily and rebuild them quickly.
Can you point out the lines in the source code to any version of the linux kernel that indicate being narrowly targeted towards serving http? Or are you just spouting crap (while blowing off the fact that Windows95 with its wide user base has a different code base than NT server with its web server revenue)?
Great enthusiasm - You sound like Steve Balmer :) The problem here is that you're wrong on several key points:
1. Win2K's interface is not improved. It sucks. NT4 was good. I get paid to admin NT4. I like NT4. Win2k is a major step backwards in usability. The ungodly number of wizards in NT5 (oops win2k) makes it impossible to do any real work. Sure you can turn them off, but the mere sight of them drives me nuts - it's like having 4000 of those fscking dancing paperclips. This is supposed to be a server os - wizards don't belong on a server os.
2. Win2k's performance. This sucks too. Win2k takes ages to boot. Once it's up, using office 2k takes far longer than NT4 + Office97 ever did. My box is a PII 450 128Mb of RAM, I know that's not enough for the 2k products, but the company won't splurge for an upgrade.
3. Stability. This is anecdotal, but I've had more lockups (5) and blue screens (1) with NT5 than I had on the same box with NT4 (3)lock and (0)BSOD - Admitedly it's still in beta
4. Ease of development. There is a special place in the most fiery pit of hell for someone who names a function RegisterServiceCtrlHandlerW() Don't tell me that Win32 makes life easier for developers. It spawns carpal tunnel is what it does.
The only great thing about Microsoft was that my manager there was damn sexy. I couldn't get any work done without a bag of ice in my pants.
--Shoeboy
What about latency? Is there an interest in not hitting a statistical percentage of your users with lag from hell or a forgotten connection? If the NT box gives _you_ the short end of the stick, do you really care that it's serving lots of other people faster than Linux would, when Linux would have got around to you sooner? Is it worth an additional quarter of a second penalty for most people to spare _you_ and others the occasional zombie act? Scheduling costs. Giving individual requests a sort of value costs if you compare it to thinking of them only as percentage points. I don't care if NT never crashed, did CGI scripts and was tweakable by VT100- if it places its own 'score' ahead of the fair distribution of load over the pending requests it's responsible for, I don't want it. It's like having a heart that pumps 200% better, 99.9% of the time, but the other 0.1% of the time it forgets to pump for a couple minutes. The priorities are totally out of whack here.
... and next time the advantage will be mine.
You think that the Linux-kernel coders rolls over and play dead? I don't think so...
You obviously know nothing of unix or NT. How is NT easier to administer than unix? I used to think so until I LEARNED unix, now I can't believe how utterly clunky NT's pathetic management tools are. Worthless.
FWIW, unix (all kinds) are an ADMINISTRATORS OS. NT is a LUSERS OS. NT is absolutely horrible to administer.
support gun control: take guns from cops
Whilst I doubt manyfolk are going to be running Netscape on ther server and browsing, this still is a valid point.
Namely, what if it was another service you wanted to run whilst serving up webpages?
What would the results be if the machines were also acting as DNS, email, or backend database servers too? Or even running CGI's? These were only static tests. I recall seeing an article PC mag did before on web server performance. The static performance was similar to this test, but IIRC, NT got tromped in the CGI tests.
-- -- The Dragon De Monsyne
Hang on a second- 'hooks' doesn't tell you much.
What I'd guess is that NT is doing some very funky caching so it can serve pages in bursts. Here's what I'd do if I was NT: save up requests in queues until I could load up a given page and whang it out to a bunch of clients without even blinking or pausing. I'd have everybody sitting around waiting for me to be happy with the size of the queue, and I'd be real busy trying to make bigger queues the better to whang out a pre-loaded page over and over as fast as I possibly could.
If I was Linux, I would try to keep track of what queues were forming, and rush about trying _not_ to let any of the lines get too big.
Considering that people want not to wait when they get web pages, doesn't it seem preferable to pay a lot more attention to latency than MS is willing to? There are genuinely conflicting interests here. To begin really obliterating the benchmarks, it becomes increasingly important to not waste much time caring about if anybody's been waiting too long- you _want_ them cooling their heels in line, that way you can gear up and do whole lines in mass-production fashion.
Lastly, picture what happens when there isn't so much of a line. Linux happily serves the customer. NT _first_ looks to see if it can make the customer get in line and wait for others like 'im...
Remember that these benchmarks still come from an organization funded by Microsoft, and an organization that has done slanted and biased benchmarks before. That doesn't mean that these results are automatically false or that they have no value; however, one should not take these benchmarks at face value without looking further at the details of the test, especially since other benchmarks have come to the opposite conclusion. ZDNet, if I remember right, did a benchmark earlier and concluded that Linux 'ate NT's lunch'. And that benchmark is not the only one that has concluded that.
Mindcraft does have a vested stake in having the results come out in NT's favor. It was embarassed previously when its original benchmarks were shown to be essentially a hatchet job funded by Microsoft. To make both Microsoft happy and salvage its own reputation, it has to
1) Produce results that still favor its backer, Microsoft, and
2) Have the results favor NT by a lower margin that the previous test so that it looks like the later tests are 'unbiased'.
I am not saying that their results are necessarily incorrect. What I'm saying is that Mindcraft has a vested interest in having the outcome of the tests come out a certain way, and may still have found a way to rig the tests. Before jumping to conclusions, we should learn the details of the testing to see if significant slant is still present in the tests.
> Last message I read about it on klm says that performance has recently matched apache.
Any Apache developer will tell you that's nothing to brag about.
I've finally had it: until slashdot gets article moderation, I am not coming back.
I'm not sure if i directly compares Win32 to KDE/Gnome, but i'll have to mention here that KDE and Gnome aren't GUIs, they are specifications and APIs for writing X apps. Certainly Gnome isn't a GUI, but rather a specification (and API set) for writing X applications.
Windows' WDM (Windows Drive Model) new, driver model is now enhanced to allow more stable systems when they go into sleep or hibernation mode (for windows 2000), MS has always allowed programmers to extends the kernel by making drovers/vxds etc, but poorly written drivers lead to poor stability. The reason this happens on NT more often is simply cause NT has more hardware and drivers than Linux, not to mention more variety and mix-match combinations.
As for personal experiences of stability with both OSs, when I use Windows 2000, it just 'feels' more stable than Linux for me, as well as feels faster, and obviously has more software that I prefer to use like Visual Studio, Office and Java2 - Word runs faster than anything I've seen on Linux that comes close to functionality.
People who try to compare sizes by comparing Emacs to Office, or Lynx to IE is just silly.
If there is such a thing as the real problem, it's not mostly with Windows, it's with Windows developers who get ahead of themselves.
I get more support from MS than any other software company, or even Linux community. I get bitten whenever i ask a programming question on a Linux newsgroup, which is why i kindda ran away back to MS - the people tend to be nicer and more tolerant. Win32 API IMHO is a well rounded API. Easy to work with things from Threads, to Pipes, and to GUI stuff, like menus etc...KDE and Gnome need to play catch up..but by the time they catch up to win95, they'll need to catch up to Win2000.
I don't have a big problem with Open Source, I have a big problem with the Open Source community who seem to be too religous for my liking.
I'd rather go with Microsoft and pay money, than with Linux and have a community who bites heads off.
This indicates that the Samba context switches may indeed be significant.
I find these studies inadequate as data to inform a purchasing decision. While MS will claim that they have proven NT to be better than Linux for Web and file serving in the general case, I disagree. Here's why:
These studies do not address price/performance. P/P is one of the most important metrics in making a purchase decision; these studies measured only peak performance. That the prices of the Linux-based and NT configurations tested are not given indicates to me that Microsoft wishes price to be disregarded as a factor in purchasing decisions. To do so would be an irresponsible act for any purchaser. Consider that NT license fees increase dramatically with number of clients, while Linux's price is constant and lower than any NT option.
These studies do not address options such as clustering. Clustering is a common solution to the problem of constant high client load. It may well be a better solution (in P/P and in peak performance terms) than simply boosting processing power with multiple processors. It also has reliability advantages.
These studies are not generalizable to other hardware configurations. While MS will claim that they prove that "NT is faster than Linux" inherently, they do not. The HW configuration was selected for the first Mindcraft study, which has been proven to have been engineered to favor Microsoft. Hence the hardware configuration itself is suspect. An across-the-board comparison on various configurations, with P/P as well as peak performance measured, would be a more reasonable comparison of the virtues of the OSes themselves, and would also highlight particular combinations of HW and SW that are worthy of consideration for purchase.
These studies do not address security. The release version of MS IIS has outstanding security holes, including the recent one disclosed by eEye. This was a root compromise which took eight days for Microsoft to admit, and two more to fix. Microsoft classically avoids the subject of real-world security, preferring the proven-worthless tactic of security by obscurity. Security, of course, is a major consideration to be made in purchasing.
These studies do not address stability. Stability, like P/P, is an important metric for purchase decisions. It helps one determine how expensive a system will be to maintain -- one that requires regular resetting or reconfiguration in order to keep operating will cost in manpower; one which crashes a lot will cost in downtime. Downtime costs money in an enterprise situation, and hence should inform purchase decisions strongly.
These studies do not address changing real-world needs. A real server system is rarely left serving static Web pages forever. When needs change, performance will likely change as well. Building a system to meet a single, narrow-minded need is likely to lead to a dead end in terms of scalability.
These studies demonstrate nothing about the future. Based on past trends, one can expect the situation for Linux-based OSes to get better and better. The next version of Windows NT will likely offer decreased performance on the same hardware (due to increased resource consumption by the OS itself) whereas future versions of Linux will likely improve performance. Buying heavily into Windows NT leads one to platform lock-in which may damage one's ability to escape the expensive effects of bloat.
In short, I do not believe that MS has demonstrated that there are advantages to purchasing an NT system over a Linux-based system for real-world file and Web service. Wise system administrators, IS/IT managers, and CIOs should stick with the proven security responsiveness, stability, price/performance, and scalability of Unix-based systems, possibly including Linux-based systems, rather than betting the farm on the Johnny-come-lately Windows NT.
Like the article says, fixes are already on the way. How quickly can Microsoft adapt to such findings? Maybe it's not Microsoft of Borg at all.......
Yes, well. Obviously you can't recompile windows without features to slim it down, but like i said previously, Windows is a general purpose OS, it's designed to do most things reasonably well. ...where MS just want to integrate IE into Windows (a very good idea IMHO). :) :(
This saying has been posted many times in this (NT vs. Linux: Again) thread.. "In my experience" Windows is unreliable. I get a minimum of 4 crashes per day on this machine (two of which I blame on the Cyrix processor but the other 2 are definitely Win faults). Even back in Win3 I got many BSOD's and I thought that was normal. (first PC was a 286 with DOS 3.3, nothing to compare with really)
I'm talking more of writing apps to expect user errors and to handle them well, which is sometimes quite hard when you are writing GUI apps. Also, MS apps have technologies such as intellisense etc which add more overhead. It's try that COM/ActiveX would be more bloated than 'raw' apps, but I see the benfits are worth it. For example, IE can view PDF files, not by implementing it itself, but by calling Adobe Acrobat and telling it where to put itself (IE as the OLE container). Other things like, being able to put a word document inside an excel application inside a bitmap, inside the windows activex desktop etc etc. Very cool.
The programs don't need to be as large as they are.. error traps are a must, wether it's made for the 'dumb user' or not. It must be the way they make the things. I don't see them as needing to be 6 megs for a simple program.. *nostalgic for the 10K kickass progs*
There are other things like consistancy. z as an undo for almost everything, including this form text box i'm typing into. These are small things, but there are hundreds of small things which makes windows so much nicer to work with.
That is nice, agreed. But any suite of programs made well will give you that (gui or otherwise) usually in a smaller package. (Don't mention Office or something like that, I know they need a platform to run on already)
That may be true to a degree, but I've never put the fear of being usurped in the way of developing anything i thought would be 'cool'. I think what you say applies to ever other industry and company tho. Speaking of other companies, do you think working at Netscape (yucky software), Sun, Oracle etc would be as 'polluting'?
There are also monetary concerns, but that's beside the point. And yea, It's not just M$ that I fear in that respect, there are many others..
Working for Sun/NS/Oracle would only be so were their mindsets so altered to not be to the benefit of all. Most of us can't go out and buy a $300 piece of software you know..
've always been impressed by Microsoft's 'relaxed' working enviroment, and Microsoft's origin as a geeky company...Gates still acts like he's 20 years old - and he's got a genuine interest in 'cool' technology like speech technology etc, and MS competition or no compettion still spend billions on R&D cause bill just loves gadegets, which is more than I can say for Ellison and McNealy (who IMHO are the real money hungry jealous types). Remember, Gates used to be a Geek, not a satan worshiper.
I do like Gates personally for those reasons actually. I used to view him as a 'model of success' until I learned some of the underhanded things he's done.. kinda burst that bubble. (I aint a youngass kid either, just some guy without much money to buy new computers, etc) I have no beef with the man himself, it's his company that's a problem.
You might think it isn't significant, but the overall impression of Microsoft I get isn't bad. Ok, I 'may be' naive but I know how easily what they say can be twisted.
That applies to anyone and anything tho doesn't it? It just happens that M$ is a very good target who quite possibly deserves everything they get..
What people see as bad things (which MS does) is nothing unique to MS, it's done worse by MS's competitors, they just don't do successfully, so it's easier for them to play 'innocent'.
Corporations/companies are human nature's bad sides amplified a hundredfold. Microsoft is currently the undisputed largest of them atm so that is to be expected. With any luck we can all tone down these powerhungry 'institutions' so we the people can go through life without being under the thumb. If M$ had been allowed to continue as they were going, Unixes would be almost dead (for all intents and purposes), Mac would be 'that other computer'.. Be would soon be bought by them if they won't already be anyway.. and the end user would be locked into 'Upgrade the computer so I can still use the Internet, or pay rent for the next 2 months'. And that sucks ass.
Netscape have used much worse strong arm tactics to try to control web content
Maybe for the computer know-naughts it would be good for them to think everything is the Internet, including the PC they shelled out 1.2G's for, but I tend to want to give people a chance to learn first. If they prove their idiocy then they can do something else. If they 'get it', even a little, then all the better for us. No need for everything to be '100% dumb-fuck-compatible'.
Netscape have always been free to create a replacement shell to Windows (it's not hard) that supports netscape web integration. It was netscape's channels bar that wouldn't let you delete default channels, not IE's, and it was Netscape that made Netcentre the default page for every browser on your system. Without asking. How rude.
rofl
Netscape is no better than M$, though they claim to be. Mozilla is a plus for us at any rate, and I think more than M$ has done.
They are all, in the end, trying to control what we see, how we see it, and in effect our lives. Things like the FSF, W3C, etc are good because they further the things that benefit all. Low techies can view HTML with W3C standards in it. Programs are simple enough to make it easy to get a grasp on it right off, but are still more than good enough for the technically minded.
BTW, what's with the title. Was that sarcasm?
I honestly haven't a clue what that means.
Later.
In fact, if you guess that a reboot takes 5 minutes (?), and you had a system to automate the reboot process, 99.9% uptime means that the machine can go down about once every 3-4 days.. so someone claiming 99.9% uptime can just bring the system down that often, to "refresh" the system.. I personally don't think that qualifies as "stable". 99.9% uptime is not impressive.
..Jeff Keegan
seven syllables explain TiVo: kee gan dot org slash ti vo
It doesn't matter if it made NT unstable.
I am telling you now that it is possible to trade off latency for overall hit volume. For the most primitive example, to really illustrate the point, consider this- say looking up a page takes X amount of time, and serving it takes Y amount. You'd think that all requests would go XY, XY, XY like that. Instead, set up a caching system thusly: all requests form into queues- and _only_ when a queue has 1000 requests does it get served, all at once like YYYYYYYYYYYYY. There is no attention paid to the queues other than this 1000 request limit.
This arrangement would beat Linux. Soundly. Severely. If given an insanely fast barrage of mixed requests, it will keep the bandwidth hopping with 1000-hit bursts, far faster than Linux could hope to approach.
Of course, if one of those pages has only 900 requests, those requests _sit_ there. Forever. _Never_ being served.
Does latency begin to make more sense now?
Linux is clearly not caching as aggressively as NT. The question that needs to be asked is this- is NT, for the purposes of these benches, pursuing a caching scheme that is unacceptable for normal use? I have illustrated a setup that will beat Linux, might even beat NT- but which is useless for realworld tasks. But there is no arguing that (properly tweaked) this arrangement will run faster, how fast depending on just how efficient the 'Y' phase of repetitively sending a page to different clients can be made. And none of this so much as mentions the fact that if you put it on the web, most people would get _no_ response from the server at all, as it would be waiting to fill up its caches before acting.
Benchmark priorities are irreconcilably different with realworld priorities...
Another question to ask about this test is how many sites handle 36253440 hits/day? Even CNN says that they only handle about 3 million hits on a normal day.
Hmm, that's an interesting theory. Somebody should test this. I'd be interested to see how Linux and NT each perform as a webserver if each has 500 T1 visitors and two 33.6 modem visitors. Will NT pump data into the T1 visitors forever, forgetting about the low-bandwidth modem users? Will Linux be better in this respect?
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Well I personally would never run the gui on a server - a gui is for client machines.
I have had X crash (it's pretty rare compared to a 95/98 crash, I haven't had a lot to do with NT tho). It will not bring down the OS the way a Windows crash will - even if it does lock your console you should still be able to telnet in and kill X - then again I've never had a problem switching VTs when X crashes so I wouldn't know. The GGI people are working on putting accelerated drivers in the Kernel and I think Matrox cards already have some acceleration on a standard 2.2 kernel.
--
enterfornone - logging in for a change
Close but not quite. I wasn't suggesting "that NT was modified to be app. specific to target a narrow purpose (web servers)...". I was suggesting that it had been extended to perform well in the static html subset of this particular, high-profile, area of service -- with an eye on benchmarks rather than, and, possibly, at the expense of, actual user needs. Perhaps that is what you meant as well.
:)
Your point that "MS has a far wider user base to afford making their OS better for just one app. at the sacrifice of other applications." only makes it a greater sin. It does not prove them innocent. Consider their slipping market share for NT in the server arena vs. Linux. This is the kind of advertising that they cannot buy. I would suggest that your logic applies here too -- can they afford to pass up such a chance to prominently display the "superiority" of NT with the disinterested blessing of ZD?
As for your annoyance "that people are coming up with all kinds of excuses instead of facing the facts", which facts did you have in mind? Note that I did not question the facts of the result, only the architectural decisions that contributed to NT winning these particular benchmarks.
On your claim that "This reveals our own denial of reality", do you know of a wide-ranging suite of disparate and concurrent benchmarks that show this NT performance advantage to be more than specific to these two areas? Please, update my reality and share them with me. If you can demonstrate that I am off-base, that these results are not anomalous, that this performance is typical of general, mixed-load, server performance, I'll humbly apologize to both you and Microsoft.
Geeky modern art T-shirts
This is absolutely a troll. Did anyone actually visit the site -- Atipa Linux Solutions?
Actually, it could be sarcasm -- saying "Yeah, sure, we're going to change our entire strategy based on this one benchmark... NOT."
--
Do I look like I speak for my employer?
X has crashed on me a number of times, cutting off the console and any ability to kill X or even safely reboot the machine. Sure, the Linux kernel wasn't at fault, it was chugging away happily. But it was at fault by negligence, because it didn't do it's job -- safely abstract the hardware.
Frame buffers are a start, though my understanding is that the X server based on the frame buffer is rather slow because it doesn't have any acceleration available to it. So it's only a start.
Posted by SmashPHASE:
:)
So if I get it right, web servers with an extreme
payload perform better with NT?
Assuming Slashdot runs Linux, let's go for NT, since it also runs faster on single cpu
configurations, their M6800 (or is it a Z80?) will probably benefit from it...
"Life's a bitch and than you marry one.."
You refuse to identify yourself or the company you claim to discuss, and you expect us to believe you.
Go back under your bridge.
--
Do I look like I speak for my employer?
Granted, if it's possible to do it in under 200 megs (my Win partition takes about 210ish of the 450 I allocated it, custom install). I don't see this in Window's future
...where MS just want to integrate IE into Windows (a very good idea IMHO). Netscape have always been free to create a replacement shell to Windows (it's not hard) that supports netscape web integration. It was netscape's channels bar that wouldn't let you delete default channels, not IE's, and it was Netscape that made Netcentre the default page for every browser on your system. Without asking. How rude.
Yes, well. Obviously you can't recompile windows without features to slim it down, but like i said previously, Windows is a general purpose OS, it's designed to do most things reasonably well.
Hmm.. so if I want to make a compiler for instance, I should make it so simplified that some shmuch who knows jack and is too fscking stupid to know it anyway can belch at the computer and churn out mind-blowing software? Come on. If someone can't do something (my example was admittedly very exaggerated) they have no business doing it. If they are learning it, have an interest, and most importantly an ability, then they are more than welcome. And let me be the last one to flame them. Unless what they do sucks
I'm talking more of writing apps to expect user errors and to handle them well, which is sometimes quite hard when you are writing GUI apps. Also, MS apps have technologies such as intellisense etc which add more overhead. It's try that COM/ActiveX would be more bloated than 'raw' apps, but I see the benfits are worth it. For example, IE can view PDF files, not by implementing it itself, but by calling Adobe Acrobat and telling it where to put itself (IE as the OLE container). Other things like, being able to put a word document inside an excel application inside a bitmap, inside the windows activex desktop etc etc. Very cool.
There are other things like consistancy. z as an undo for almost everything, including this form text box i'm typing into. These are small things, but there are hundreds of small things which makes windows so much nicer to work with.
Nope, by no means. Working for Microsoft pollutes you. To varying degrees, I should have said. M$ and their ilk (other companies with similar policies) drag down those who would be giants in other aspects than the dollar sign. I am somewhat fearful, truth be known, to develop my better ideas due to Microsoft and others like it. I fear my ideas would get stolen right out from under me and there would be a total of jack SHIT I could do to retaliate
That may be true to a degree, but I've never put the fear of being usurped in the way of developing anything i thought would be 'cool'. I think what you say applies to ever other industry and company tho. Speaking of other companies, do you think working at Netscape (yucky software), Sun, Oracle etc would be as 'polluting'?
I've always been impressed by Microsoft's 'relaxed' working enviroment, and Microsoft's origin as a geeky company...Gates still acts like he's 20 years old - and he's got a genuine interest in 'cool' technology like speech technology etc, and MS competition or no compettion still spend billions on R&D cause bill just loves gadegets, which is more than I can say for Ellison and McNealy (who IMHO are the real money hungry jealous types). Remember, Gates used to be a Geek, not a satan worshiper.
You might think it isn't significant, but the overall impression of Microsoft I get isn't bad. Ok, I 'may be' naive but I know how easily what they say can be twisted.
A little MS employee sending an email saying "Netscape navigator is quite popular" could snowball into an accusation that Microsoft is trying to take over the world and that they are working for an alien goverment.
What people see as bad things (which MS does) is nothing unique to MS, it's done worse by MS's competitors, they just don't do successfully, so it's easier for them to play 'innocent'.
Netscape have used much worse strong arm tactics to try to control web content
BTW, what's with the title. Was that sarcasm?
Talk about troll.
Ok. If you insist.
Replace Microsoft with Linux Torvalds.
Hello? Anyone home inside there pal? He doesn't make the GNU tools, nor Apache, or any of that. He makes the kernel. Go read before you open your mouth and say something stupid.
Linux doesn't run SMPs well does it?
NT isn't designed to be a super computer OS. It's a PC operating system, a general purpose OS. DUH.
It's a PC OS huh? CP/M was too correct? General purpose? Yeah right. I could list things that will not run on NT (win progs) but I won't make you look any dumber. You do good enough. DUH.
slashdot.org doesn't run IIS? uh..it's got nothing to do with microsoft..so what?
If you don't see the implication there you need help...
EBAY, Microsoft, Dell run IIS, and they have much bigger websites than slashdot.
Yeah, I never looked at Ebay, and hate Dell. As for the M$ site, every *every* HTTP request I send their servers returns 'Remote connection reset by peer' in Netscape. Nice server.
Microsoft don't grow engineers on trees, their engineers come fromvarious backgrounds (inlcuding unix). They have enough money to hire the best in the world, and they do.
Yeah, you'd think at a cerain point there's such a thing as ENOUGH money.. And when they hire a programmer, he may be creative, smart, innovative, all that crap. But he is no longer 'pure'. He prolly expects to get paid when he goes out with his wife for his 'service' (dinner, not sex).
You probably have your face stuck up somewhere dark to realise you can't comapre vi or emacs to Office 2000 and complain how large Office is etc. Office does MUCH more, and Microsoft's products simplyfy working, which is more than I can say for Linux/Unix.
All M$ products are overly bloated for one thing, and Office is no exception. Sure it does lots of kewl little things, but hell, I can make a picture that does lots of kewl things with two pencils and some resin (from a tree).
They simplify working by making everyone work the way THEY want them to. Nice company.
Sure, there are you guys out there who don't want things to be simple, you'd rather excercise your brains doing "hard" things like mounting NFS/SMB dirves by typing rather than doing it in a few clicks.
You go ahead and play with your mouse. We know you depend on that little thing. We however know, and will continue to, how to do things without a mouse. Guess who's gonna be using who's programs here?
I prefer to have the OS do as much as it can, while I get on with the real work. If by any chance, I need to do things manually, I go and do it.
Oh, Win does as much as it can. Mostly collecting files it doesn't need, eating your prefs/settings, and if you are really lucky it might eat a partition or two. Nice OS.
And what's your problem? Are you on medication?
Why, you got something good?
MS Write, MS Bob? So what? How about MS Windows, MS Office (Word, Excel, Access, Powerpoint etc), MS Visual Studio, MS J++ (if the best selling javaproduct), MS Exchange, MS SQL Server, MS Internet Explorer, MS IIS, MS COM (the most successful component model in the entire world), MS MTS, MS DTC..all pretty much defacto standards now...and that's only to mention a few.
Those are standards (well, the ones that are in that list) are only because of brute force and M$'s anticompetitive nature. How can you compete with 100 bucks in your pocket when they got a billion they'd just as soon stick up your ass as anything?
Unlike Linux users MS doesn't claim not to make mistakes, infact Gates even showed the video of Win98's BSOD last year, again this year at COMDEX.
So you compare Linux *USERS* to Microsoft's *PROGRAMMERS* eh? You think every Joe who uses Linux is a programmer? I pity you and your world.
That video is something I'd like to see again tho.. always good for a laugh. Although what's the point of reshowing a BSOD, truthfully? Who hasn't seen more than they can possibly count of them already? Kinda redundant if you ask me.
Happy clicking. I'll be off to play around with my Linux box, to change it's basic settings. Like to see NT (95/98/2K) do that. *chuckle*
1) I think this shows that both Linux and NT are insanely fast under these circumstances. Both the setups in this benchmark pumped out enough data to satisfy the needs of nearly any possible webserver or fileserver you can think of. What type of server gets more than 150 million hits a day?
2) Why isn't BSD ever tested in these benchmarks? Many sites, such as cdrom.com, put up impressive numbers with BSD systems. It'd at least be nice to see what BSD can do with a 4-CPU/4-ethernet card box.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
A test I would like to see now is a "real world" test. Using the same techniques as in these tests, but set a price cap on all tested items. Say the cap was 10,000$ or so. This is alot more realistic because this is how real companies operate, they are on a budget and can't always afford 25k worth of hardware and software. This 10k would have to be used for hardware AND software at fair market value, the same price a business has to pay for it. If you could only afford a dual processor NT system and a quad processor linux system then so be it. Besides a price cap, how about a longer testing period, say about a week and a half under constant high loads. If the system crashes you can reboot it, but the final results will be a total of pages served in the given time frame. A test like this would see which system would give you the biggest bang for your buck, if your web server or file server crashes and is down for a long time then you're going to lose money, every page not viewed is a dollar not going into your wallet and your file server is usually critical to your company, if your people can't view and work with the documents they need on the server then you're going to lose alot of productivity.
The price cap would severly limit the power of your hardware but it would do so within reasonable levels. The test would be less of a pure performance test and more of a price/performance test. This is where the free unicies excel, they are excellent operating systems yet cost you nothing more than download time or the price of media. Maybe other free unicies should be thrown into the mix, which would give a little more variety in the comparisons. That kind of test would turn my head if I were a sys admin, or even the owner of a business where every dollar counts and if I'm wasting my money on NT because it has a better marketing pitch I can change my ways the next hardware upgrade or just replace NT with linux on my servers.
I'm a loner Dottie, a Rebel.
that doesn't explain the HTTP perf diff though... Is IIS also in the kernel on NT?
Consider who you're talking about; I'd consider that while IIS itself probably isn't in the kernel, it access top-secret M$ stuff which is.
also, why can't we do a similar multi-threaded implementation on Linux?
I don't know. It probably is possible; it just hasn't been done yet. I'd consider these tests to be a sign that it needs doing.
That's what M$ did when they found problems with NT file throughput from servers to clients. That's what Apple is doing with their OS, finally bringing it into the 21st century. Competition is what makes good tennis players, people, and it works the same way with programmers. I've no doubt that the next version of Linux will whup ass on NT, the makers of which will then come back with something that whups ass on Linux, and then the Amiga will come back huge and crush *everybody* ;-).
I didn't say the linux kernel was narrowly targeted. I was referring to Sun Tzu's comment:
"Realisticly, what I think MS has done here is create a "benchmark special". They have picked two high-profile applications and integrated them into the kernel a little too intimately so they can claim that NT in general is faster than Linux."
If you think about it, it's quite unlikely that MS would deliberately create a benchmark special OS to beat linux. It would be possible if web servers were their primary business and they wouldn't weaken the rest of their client base by specifically modifying the NT kernel just for web servers. However, MS has such a huge market that this would be a stupid move, esp. since they have been striving to merge the 95/NT code bases and haven't been able to do so yet.
Why would they jeopardize that goal and risk screwing up software for their 200 million paying customers just to satisfy a benchmark against linux for a webserver that they give away for free?
L.
L.
And as the improvements flow from the Open Source groups as a result of these tests it will only get better for those of us that have already switched over.
Peace
--Somewhere there is a village missing an idiot.
I think this is the first time anything like this has happened.
None of you are crying fowl.
I am very happy to see that the Linux community can see when they have to stop whining, and start programming.
I think Linux has a lot of potential as a enterprise level server. But that's all it is. Potential. Without people willing to take the time to make use of that potential, Linux will die.
I can't wait to see the rematch. And I congratulate NT on it's victory.
(this is more petty bickering than anything worth reading)
NT has preemptive multithreading, and IIS uses NT's threading to do it's work (including getting requests, and assigning them to a thread). You won't see a problem with what you're talking about. That's the whole point of threading, and why NT is faster than Linux.
Excuse me, but unless you're running apache from inetd (which is stupid for ANY amount of traffic), apache uses the Linux kernel to accept connections, and passes the connections to its child processes (subservers), which takes advantage of Linux's preemptive multitasking. This is exactly what you were claiming NT does that makes it better than linux. And if the question is one of multitasking vs. multithreading, I don't see why moving all the children into a single multithreaded process would add any performance gain, and would make the apache code much more complicated, because it would have to deal with mutexes and locks and all that un-fun stuff.
It's more likely that Linux is guilty of what you say (certainly, we know that the ip stack of linux is guilty of this). And the performance of linux on SMPs shows how guilty it is of pending other tasks because of it's lack of threading (or use of).
Linux has a much lower per-process overhead than NT does, therefore making multithreading within the same process merely an alternative to multi-process multitasking (which was WRITTEN for this kind of thing, and places no constraints on explicit process code to take advantage of it) as opposed to a necessity (as is the case in NT).
--
Paranoid
Paranoid
Bwaahahahahaa.
As much as I hate Microsoft for what they've done to the computer industry, I have to admit they do have the ability to buy some good engineers. However, I think you're missing two important points regarding the relative programming talent in the Microsoft and Linux camps. First, there are plenty of terrible programmers at Microsoft as well. Steve Maguire's book Writing Solid Code alludes very subtly to this fact. (They grew so fast and there was such a shortage of qualified talent they were forced to hire less talented people.) Second, there are lots of excellent engineers working on Linux. In fact your implicit assertion that there are only "average hackers" working on Linux is rather absurd.
Everybody knows the rock stars of Linux development, like Linus and Alan, but I think there is something even greater brewing. From personal experience and contacts it seems that there is a lot of anti-Microsoft and pro-Linux sentiment in computer science departments all over the world - there certainly is in mine. I think most of the very best programmers would work for less money for the satisfaction of doing something they believe in. With Linux they can really feel like they're making a positive difference. At Microsoft the only real motivation seems to be money - they certainly don't seem to mind selling poor products to consumers.
What I'm saying is that Linux has won the "hearts and minds" battle, so to speak, for recruiting engineering talent, and the money is starting to follow as well. I think the imapct of this over the next few years will blow everyone away - especially Microsoft!
Chris
simpkins@tilc.com
(BTW, the links above to simpkins.org don't work right now. I'm moving my stuff to a new server.)
>Second, I am not talking about the Mindcraft tests. I am talking about
Dude, the PC Week tests *WERE* the Mindcraft test done by people who seemed to know what they were doing.
Didn't you notice that the numbers doesn't seem to be as out-of-wack as the orginal test performed by Mindcraft with no one looking over their shoulders? That speaks volumes about the basic compentence of the people at Mindcraft to perform this kind of testing.
> The world seems (from this side at least) to speak primarily english.
;)
For this side yes. It seems to me that mandarin (or maybe another Chinese language) is the most used language. hey, 1 billion people that makes one out of six.
Go everyone and learn Chinese
"The obvious mathematical breakthrough would be development of an easy way to factor large prime numbers." Bill Gates,
First two mistakes I made in my first post. It should be 362534400 hits/day, and 3 million pages serverd per day.
Hm, if you look, there are two posts by this AC which are EXACTLY the same. There is an interesting thread on the first one. This one is um... redundant.
Also, you must take into account stability. From what I understand the NT tweaks made the system very unstable. There are plenty of things you could do to linux (since you have the source) to acheive the same tradeoff. Also, there are already patches to fix some of the problems found in the original mindcraft test. They have been around for along time. Just the linux team wasn't allowed to use any patches after apr 15 (or some date close to that). I would like to see how linux does compared to nt w/ the new patches. I think this is the true power of linux. The ability to say "yah work does need to be done in that area" and being able to get the patches in a few weeks, not months, or just writing them yourself.
-matt
There are a few replies that I probably could surmize have been written by the microsoft camp, and I guess this could be beneficial to wake up some of you coders out there. I look at all these reports in this light, nothing is ever perfect but things can always get better, NT needs to, and so does Linux. Now you all can argue to the degree to which each OS needs to improve, but I think the thing you have to realize, the enemy is our own pride, the competition, is microsoft, this is how we have to see it.
Invision a new reality, a new world
with GNU/Linux Debian Rocks!!! =) NiteRain!!!
I don't know what life is, but no one gets out alive...N
You're saying that Linux + 4-way beats NT + uniprocessor in price?
First thing.... These tests where much better but they still manage to miss the mark.
.htaccess files on commonly accessed files nested five directories deep. Not .htaccess file .htaccess files as possible. This
.htaccess file that is already slowing things down also has
:)
Ok, just in case anyone still thinks these tests are worth a shit. I'd like to clearify that this is pure and unadulterated shit. There now that the
childish remarks are through. I'll do some intelligent speaking.
First off, I don't doubt this to be shit from the get go. I'm an MCSE (my work paid for it) and I know the insane amount of system reasources it
takes to run a NT Server alone. Yes, I know how to properly configure an NT Server right down to the streamlining of the registry. Plus, we have
all been through the multiple restarts and memory that applications won't let go of after using it. Not to mention all the swapping and overhead
processing. Don't get me started with IIS 4.0.
There is a new bug found almost on a daily basis that spells doom for these servers. Plus, IIS 4.0 doesn't have near the amount of features and
configuration possiblities as Apache does. Next Apache needs someone who knows it inside and out to configure it. This is due to Apache's
extreme flexiablity.
Say that average joe smith sets up his Apache server and uses
uncommon with big sites where management is broken up. Well, for every request on the document Apache will check with each
per directory. So if this file is accessed 100 times. Apache will check 500 times for the rights to that file.
Because it will check the root to the next directory to the next. And merge the config files it finds along the way. Making Apache check 5 times
per document requested. But, on the up side if you need infinately specific rights to files. This is a god send that can be reduced by placing
commonly requested documents near the root of the (don't fork the directories too much)server. And using as few
is why you should try to place as much configuration as possible in the gobal configuration files and preferrably in the server configuration file. I'll
explain the last part of that last sentence next.
When Apache is looking into what the rights are for a requested file. It checks certain files in certain orders. And within those files it checks it
against the directives in the order they are placed in the config file. Meaning if that same
the most requested file in the directory near the bottom of the config file. It will take longer. Maybe not whole seconds longer. But, enough on
heavy sites to make an impact.
These are just two of the many configuration tips for Apache a person can pick up when they rtfm (Read The Fucking Manual)and even reading
the source.
And all the rest of the way IIS doesn't have as flexiable a rights system. Nor does it handle dynamic pages as well as Apache. Infact IIS 4.0 will
work fine if it isn't that compilcated a site, the pages are static, and the machine is so big it won't ever see a processor load near 100%.
Apache has that complete control rights system. It handles dynamic pages bueatifully. And doesn't freak when heavy loads hit. It will just keeps
chugging away.
As for file serving? I can't say. I'm not anywhere near an expert at samba. But, I do know that my Linux box boots faster, handles heavier loads
better, and memory management is bueatiful. And to make another remark.
RedHat should not be the version of Linux they are pitting against NT. Sorry, this isn't a direct RedHat sucks type deal. It's a use Slackware or
something and so you can minimize the system to do only what it is suppose to do. And recompile everything to be optimized with the systems
hardware. Maybe not even Slackware. Just something streamlined. Redhat is actually a great system for the home user. That's the way they
seem to be heading nowadays. And I applaud them for it. My it's now easy enough for my mother to use it.
Personally, once again you can look at the source of the tests and wonder why the outcome is the same. These companies are heavily dependent
on Microsoft products. And some have been funded by Microsoft. Mindcraft even did there tests in Microsoft's labs. Of course they aren't going
to say anything bad about Microsoft.
The real test should be here is X amount of dollars. Put, together the best system you can. Linux would kick the fucking shit out of MS. For the
amount of the software alone you could put together a Beawolf cluster that would crush any NT Enterprise 4-way SMP box. I know, I tried this
before when installing many NT systems to upgrade a hospitial. Personally, I won't go there if I'm shot. But, that Linux cluster is up to this very
day without a reboot performing critical storage and access control for CAT scan images. On the other hand the NT clusters (if you can call it
true clustering) are constantly having parts of them rebooted.
Whatever, don't believe this stuff. It's just FUD and the media looking for conflict.
Eros -- I know what every file on my box is there for..... Do you?
"the margin of victory was smaller than in Mindcraft's tests."
"Where kernel problems were found, fixes are **already** under way"
It's not FUD, but valuable advertising for Microsoft to use in their campaign against Linux. This is where REDHAT and CALDERA and SUSE
*******should*******
stand up and put out counter-advertising that plays UP the strengths of the OS instead of allowing Microsoft to run fullpage ads in PCWeek and other 'management'read magazines.
the gu ru
How many high volume web sites use static content?
And how many lesser volume web sites actually require machines with that much horsepower?
In the real world, NT loses because
1. It's way more expensive, both in inital cost and in maintainence.
2. It's a crappy environment for doing any CGI work that isn't ASP. Hell, it's a crappy environment for doing programming period.
3. IIS doesn't conform to the HTTP standard. (I've personally been screwed by it's bug of not sending cookie headers on relocates)
4. Lots of security flaws, in NT and IIS.
Okay, NT was tested to be faster. But don't you think it is hilarious that MS deems Linux as a competitor? I mean, sheesh, Linux is free! It's like the tortoise and the hare. The hare obviously has all the advantages (here is would be money, man-power, equipment, time, etc.) and the tortoise has still managed to keep the hare on it's toes. I'm not saying that Linux a direct comparison to a tortoise , no, I'm just saying that Linux is obviously a sleeping giant. MS better try harder than ever because very soon that giant will wake up.
One particular NT 3.51 configuration achieved a C3 rating (certainly not B2) when it was not connected to any network.
NT has preemptive multithreading, and IIS uses NT's threading to do it's work (including getting requests, and assigning them to a thread). You won't see a problem with what you're talking about. That's the whole point of threading, and why NT is faster than Linux. It doesn't matter how big a pipe a certain user has, NT will assign a 'timeslot' for that thread, then move onto the next, all threads will get equal priority.
It's more likely that Linux is guilty of what you say (certainly, we know that the ip stack of linux is guilty of this). And the performance of linux on SMPs shows how guilty it is of pending other tasks because of it's lack of threading (or use of).
are there features from NT that make it inherently faster than
Linux in these tests?
You find the machine which NT works best on and then
devise a set of tests to only cover the things NT is good at
(and specifically anything it is poor at.)
I'm sure that everyone out there knows someone who is a mac person. You know the type, they think that Macs are king and that everything else sucks. Well the macs were better in many ways at one time, but these guys still strut around like it was still true. They do almost as much to hurt the platform as the stooges at Apple itself. When I see people blasting the mindcraft study it reminds me of this a great deal. If we want linux to win it will take more than bitching and bragging about advantages that are no longer there. Apparently the service patches for NT are not just bugfixes. Microsoft has done some work to actually improve their product, and the linux community needs to do the same. If NT is able to beat linux today we need to find out why so that the same won't be true tomorrow. Chanting mantras like a Macolyte won't cut it. Benchmarks like this are a very good thing for linux because they spur developers to fix problems that might otherwise not have been seen. If we respond to this challenge by improving linux then in the end we will be the ones who gain from it, not Microsoft.
Muslim community leaders warn of backlash from tomorrow morning's terrorist attack.
are money the most important think in the world?
for business: yes!
but another question:
is business the most important think in the world?
i do not think so. maybe some people think it is, but they stop when this business brings its results soon: poluted air, poluted water, devastated coutry, wiped-out whole kinds of plants and animals, etc.
in such situation, "here-you-have-blank-check" policy (HYHBCP) does not work anymore because some things can't be bought!
in long term, HYHBCP is not good. to be more precise, it leads all of us to death and starving! (including the source of check).
think about it!
hany
Every once in a while, there will be a student who can help out (such as the guy who wrote The Linux Users' Guide -- he graduated from my high school a while back), but schools can't depend on people like that. ``A couple of good IT administrators'' are kind of hard to find.
no he's not. he's funny, and it's pretty much true. there has been 3 years of claims that Linux is the best flavor of icecream.
How do you force a CPU to handle an individual NIC? I know how to set the processor affinity in user mode, but how do I do this trick? (I ask because I want to actually do it). Thanks
Well... I was wrong with my first comment.
Some of you still feel the need to whine.
Next time, I will refresh the window before I post my comments.
Perhaps some of you will grow up eventually.
AMEN
http://www.bullnet.com
IIS is not in the kernel on NT, but NT does have some very specific system calls that are great for serving static webpages.
And the Linux developers are actively working on deserializing critical paths. 2.4 should be a lot better in this sense.
And even the results are significant. Using the car analogy again: "The Mercedes has a top speed of 160, while the Linuxmobile could only reach 120 (after we connected all of the spark plug wires which were inadvertently missed in our test in the last issue), but both had excellent slalom times and, in everyday driving, nobody goes 160 anyway."
I wouldn't expect Linux to come out on top anyway. The bottom line is that on the same hardware, one program is going to significantly outperform the other only if one OS does something stupid to get in the way. Now I wouldn't be totally shocked at any future stupidity in Windows. I'm convinced now that MS product development is no longer driven by end user needs (was it ever?), but by competitive tactics. If a future version decides to fork a new copy of IE every time you open a file, or won't connect to the internet without going through some "value-added" MSN portal, or won't install software that hasn't been signed by Microsoft, then we have an opportunity to pass them.
Even now, you can probably find enough MS stupidity to concoct a set of benchmarks that Linux will surely win. (I'm sure that MS funded the original Mindcraft tests only when it was pretty confident it had removed most of the stupidity in IIS.) How about benchmarking some desktop functions against an Active Desktop-hobbled Win98 box?
i read CT magazine regulary and i think CT is a magazine with lots of technical competence, and not a marketing magazine.. those guys usualy know what they are talking about. besides technical competence the sociological impacts of computers are also often covered very good. given this it is a surprise that they are not more against microsoft. i would say that they treat microsoft more then fair.
in the last issue they also had a linux vs. NT web-server shootout. they used a siemens primergy 870 maschine (4 xeon 450 cpu's, 2GB ram, intel etherpro 100 nic, mylex dac 960 raid controler, price of the maschine: about 100 000 DEM ) here the some of the results:
serving one static html page 4k size: NT and linux almost on par (linux ahead a few %) both systems answer 900 requests/s when hit with 512 concurent client process.
with 8k size static page: linux is between about 5 and 10% ahead of NT.. at 512 client processes the linux maschine serves about 600 requests/s the NT maschine about 550.
using a 4K page but selecting one random page out of 10E4 pages linux has about 830 req/sec and NT about 720. the linux line seems saturated where the NT line is bended down already..
random 4K static page out of 10E6 different pages: linux about 270 req/s while NT has never more then about 30 req/s. that means linux is some 900% ahead..
now some dynamic pages. they used plain old CGI scripts with perl. no PHP or ASP. using all 4 CPUs linux answers 210 till 250 request/s while NT is around 60!
same as above but using only 1 of the CPUs of the maschine: linux around 100 req/sec, NT around 25 req/sec.
if the script contains a sleep(3) at the beginning (to simulate slow database connection or slow client connections) the results are: linux increases the number of requests linear with the number of requesting processes and reaches about 80 req/s for 250 simulated clients. NT is saturated at around 7 req/sec. (in words: seven).
the only time that NT is ahead of linux (about a factor of 2) is when using 2 NIC cards instead of 1.
my interpretation of all this: i guess there are very few webservers where one would needs more then 100Mbit/s.. and then one would be propably better of with 2 cheaper systems doing load balancing.. given the extra reliabilty, remote managemnt, etc of linux and the better extraordinary better performance in most tests linux is the clear winner.
greetings from vienna, austria.
der mond.
Are you out of your mind man? RedHat uses the same kernel everyone else uses, so it wouldn't matter jack squat. Grow up.
>>I think that the file server portion of the test is inherently flawed in that it uses a non-native protocol under linux...SMB. SMB is native to all windows platforms. They should be testing NFS as well, then I would think that linux would fair better
:| SMP is a weakness which is not going away very soon, or very easily. It's very easy for people to say things like the above - but it takes a huge amount of effort. Linux was designed from the ground up to be a single-CPU system, and going back and unraveling all those dependencies as well as staying stable is one hell of an undertaking. Expect a couple of years before this is finished. SVR4 MP was designed from the ground up to run on multiprocessors, and indeed SVR4 based systems have excellent scalability. (check out SVR3 - SCO had to re-engineer it from inside out in order to make it anywhere near scalable). I believe that Linux should be relegated to single-CPU boxes... too many scalability and stability probs. Get a commercial Unix if you're already into investing in a multi-CPU box.
No, the testing was for Windows connectivity - NFS is not very relevant in that context. If you want to compete with NT, do it on _their_ turf.
>> IIS 4.0 vs apache was much the same, test apache vs apache
Same principle applies. We're talking about the high-profile solutions from the Windows universe and Linux universe.
>> I'd like to see NT's wolfpack clustering compared to that of Beowolf...but I guess that's not necessary considering that NT is listed nowhere in the top 500
These are completely differently-targeted clustering solutions. And schlepping together 200 Intel boxes with PVM does _not_ make Linux superior to NT in any performance scale.
>> Let's take the cue from the IP stack problem and from BeOS and multi-thread the hell out of the kernel
Yes sir, right away
lol, you got BSODs in Win31? I'd love to see that. I don't ever recall seeing any crashes in Win31.
:)
;) As soon as you 'customize' to make things work *ahem* that aren't included in the distro you'll start getting BSOD's. Be patient, it's like time. It's inevitable (sic) that it will happen, just like when the sun comes up it goes down again eventually.
:).
:(
:) NS was the fastest browser for quite awhile, and I've tried others.. still using it tho. I've never actually used Mozilla per se.. haven't even glanced at the source either (my 56K connects me at 24K, damned winmodem), too large.
:/
;D
2 09&threshold=-1&comment sort=1&mode=thread&cid=316 :/
It didn't happen as often as I may have implied (once a week or so) but the cause was usually a MS product. Go figure.
I'm running Windows 2000 beta, and I never get any crashes, but I'm being good and not installing any 3rd party hardware drivers (but i am running all the software i usually run).
Those 3rd party drivers are what makes newer hardware/software run.. if W2K has all the latest stuff in it, does that not constitute excessive bloat?
I think the point is the PC is becoming (already is?) an appliance. People don't need to know how a radio, vcr, microwave etc work, they just use it - that's the direction microsoft has been heading.
That is a dangerous place to be actually.. there will _always_ be the curious who will tinker, but the more mysterious a technology gets the less people can do with it, and the fewer people there are who can fix it if it breaks. I don't want my children, or theirs, to be stuck in a rut of relying on one gigantic company to do everything for them, do you?
Tho that attitude changes somewhat with Windows 2000, but Windows CE is their new consumer baby for d-f-compliance
I've never used a CE device so I can't really say much about that.
Mozilla isn't really a brilliant gift from netscape. Netscape give away this monstrosity of a product to the open source community. They look at it, and decide to rewrite it all.
rofl
Like they say, there's more ways than one to get there.
What's more, Netscape's object model is directly based on MS's COM. There's not much to thank netscape for....maybe MS deserves more thanks since nxCOM is derived from COM.
Thanks nothing, even tho I still use NS Comm. it crashes at least twice a day, and if I'm unlucky it takes the rest of the OS with it.
Thank the whole lot with a nice swift kick in the ass maybe..
Regarding the title, you said you agree microsoft makes great software - was that sarcasm? I'm thinking perhaps *not* now, but I could be mistaken
I didn't put the title in, my first on this thread was a Re: to here
http://slashdot.org/comments.pl?sid=99/06/25/2133
It was a response to your response to the guy who thinks MS is the best thing since sliced bread.
Later.
Sorry, you've just blown your credibility entirely. I've been over this many times. I've been illustrating a caching strategy that is guaranteed to beat linux- it may or may not be what the tested version of NT is using, but it is _orthoganal_ to CPU scheduling issues and threading. It's very amusing that you get a 'score: 2 Insightful' on this comment- methinks I sense self-effacing geeks trying to be humble at all costs here.
What I'm talking about is a strategy that's flat guaranteed to produce faster throughput- at the cost of totally obliterating latency and potentially never _choosing_ to serve certain pages, regardless of CPU usage. This _is_ a problem if you're not running benchmarks, and Linux+Apache uses a much more expensive strategy which will never be as fast as this 'caching from hell', or drop connections as willingly.
Frankly, you should have stopped with the 'get equal priority', because your continuing only illustrated that you have no idea what's being talked about. This is not a strategy for CPU utilization- that's orthoganal. It's a strategy for choosing which requests to serve and which requests to ignore for the time being- and if this is so hard to understand, I imagine the tested version of NT _is_ doing this because (a) Microsoft people are _not_ stupid, and (b) Microsoft people will always cheat given the opportunity. This is a pyrrhic cheat- you can't use it on a real web server. It has nothing to do with CPU scheduling and is purely a hack to optimize benchmarks for intranet requests.
And you, TummyX, are attempting to mislead, but probably you simply don't understand enough to realize it- you are behaving as if little CPU scheduling/threading issues have profound global effects that overcome choice of algorithm. I could specify an algorithm for NT that (given an impossibly fast load of requests) ignored everything but the requests for _one_ page, and never tried to do anything but spit out that page to the requesters, trashing all other requests quickly and efficiently. This is likely to be a simpler task than keeping track of what's _really_ happening- and as long as that one page's requests alone can come fast enough, this algorithm would beat both Linux and NT- serving only one page as fast as it possibly could and wasting no time keeping track of anything else.
Of course, this is useless- but it would win.
Now, are you going to come back and (as your interesting posting history suggests) assert again that there is no such thing as algorithms and NT is just a large vat which you throw requests at and fish web pages out of- and that CPU scheduling is the only consideration in doing this, because the only algorithm that exists is serve-upon-request?
If so, you're not only a fool, but you have no idea when you've been spanked. Please illustrate exactly when the ip stack of Linux sits around caching http requests instead of serving them, as you claim...
I hear OS/9 is very popular in embedded applications, like controlling traffic lights.
LJS
I'm an administrator for a large (1000+ node) NT 4.0 SP3 network. I have direct responsibility for about 120 workstations and 2 servers. I can say from personal experience that NT is very unstable, and it doesn't even have anything to do with the GUI-kernel connection. Our main server, a BDC, file, and print server crashed about every two weeks for a year before we finally gave up and set the damn thing to reboot every night. The thing would be sitting there with its monitor off and users would start complaining that the file server was unavailable. We'd turn on the monitor and sure enough, the machine was locked solid - only a power down would reboot it. And when the machine comes back up all you can do is look at the log files to see about when it crashed. Other than that you get no debugging info at all.
By contrast, we have a Linux box running our very active intranet web site. We've had it up for 6 months and it has run flawlessly. Interestingly, I set up the Linux web server in the first place because I was tired of IIS failing for no apparent reason (the site had been hosted by IIS).
Oh, and the Linux box is an old P166 with 16MB RAM, the NT server is a brand new Dell Poweredge 2300 dual PII 350 with 128MB RAM, hardware RAID 5, 3 hot-swappable 9 GB cheetahs. All that reliability hardware wasted on an OS that can't stay up for two weeks!
It's certainly not for technical reasons that people choose NT over *nix.
Chris
simpkins@tilc.com
(The links to simpkins.org don't work - I'm moving to a new server.)
Looking at the Microware website, it looks popular for embedded devices. I seem to remember some fairly recent video game system ran off OS-9 as well...?
Microkernels still probably makes some sense for realtime OSes. Maximum speed efficiency doesn't seem to be the primary goal on those platforms, but rather consistency and being generally lean.
Can't reproduce it here. I'm guessing it's your video card driver or your mouse driver.
I recognize that it would be nice for Linux to thrash NT on a top-of-the line box, but there is one place that Linux will definitely beat NT. How about a low-end box (eg earlier Pentium or such)? I'm using a 486 that I pulled from the trash heap as a webserver, DNS server, nameserver, et al. It was actually serving 8-10 thousand hits/day for a while on a 3C509. Not too shabby. I think this is one place that Linux really shines--setting up departmental servers or low-usage webservers for smaller companies, and you can't even run NT on the same hardware.
In the Mindcraft tests, both the NT and Linux boxes were run
quadruple-barreled with four ethernet cards. NT has a way
to bind a CPU to a particular NIC, so on a 4-CPU machine,
one CPU can be tasked exclusively to each NIC.
As I have said before this makes the NT setup half way to
being a cluster. Also IIRC this is a quite recent innovation
in NT, (SP4 maybe.)
1 - that was just an example of when doing things the long way isn't the best way. i prefer to use a keyboard than punch cards. I prefer to use a mouse to start a program (sometimes) than type in the name (unless i happen to be in a command prompt already ofcourse) :P
2 - I wasn't talking about system administration! You'd use active directory or windows scripting for that
I've only used Linux sparingly for the past 8-12 months (admittedly trying to learn it in my free time) and I've had quite a few kernal panics (5-7). I've had NT since 3.5 (1994) and have only had 5-10 BSOD (mostly due to bad video drivers for my second video card).
Everyone's experience is a little different. The presentation just depends on how we want others to perceive it.
This saying has been posted many times in this (NT vs. Linux: Again) thread.. "In my experience" Windows is unreliable. I get a minimum of 4 crashes per day on this machine (two of which I blame on the Cyrix processor but the other 2 are definitely Win faults). Even back in Win3 I got many BSOD's and I thought that was normal. (first PC was a 286 with DOS 3.3, nothing to compare with really) :)
;) I think the point is the PC is becoming (already is?) an appliance. People don't need to know how a radio, vcr, microwave etc work, they just use it - that's the direction microsoft has been heading. Tho that attitude changes somewhat with Windows 2000, but Windows CE is their new consumer baby for d-f-compliance :). Mozilla isn't really a brilliant gift from netscape. Netscape give away this monstrosity of a product to the open source community. They look at it, and decide to rewrite it all. What's more, Netscape's object model is directly based on MS's COM. There's not much to thank netscape for....maybe MS deserves more thanks since nxCOM is derived from COM..hehehehe :PPPP Regarding the title, you said you agree microsoft makes great software - was that sarcasm? I'm thinking perhaps *not* now, but I could be mistaken ;D
lol, you got BSODs in Win31? I'd love to see that. I don't ever recall seeing any crashes in Win31. I'm running Windows 2000 beta, and I never get any crashes, but I'm being good and not installing any 3rd party hardware drivers (but i am running all the software i usually run).
Maybe for the computer know-naughts it would be good for them to think everything is the Internet, including the PC they shelled out 1.2G's for, but I tend to want to give people a chance to learn first. If they prove their idiocy then they can do something else. If they 'get it', even a little, then all the better for us. No need for everything to be '100% dumb-fuck-compatible'.
lol
No, NT was modified to be app.specific to serve as many static pages as quickly as possible without regard to latency. It is possible that the tested version was in fact pushed so far in this direction that, if you made a single request for a page otherwise unrequested, it would be ignored and not serve the page.
_THAT_ is what I, at least, am implying. Why? Because I spelled out the algorithm capable of defeating normal web servers on these terms- and if you go by only bandwidth, 'mad caching' wins very big. As long as you can keep the requests coming and the granularity isn't so big as to have the server waiting for more requests too often, it pays to _not_ serve pages until you have a whole raft of identical ones to whack out in a loop. This strategy of not serving pages when asked is of course insanity for _real_ use, but by God is it a winning strategy for benchmarks! Instead of keeping all the pages in RAM, or going through a complicated prioritizing system and trying to keep latency down, you ignore all that and just count requests- and when you serve, you just do a tight loop using the same data over and over. It's beautiful, even if it is pretty useless for doing anything other than winning benchmarks- in normal use most requests will be delayed, because they're being saved up for optimal serving at a later time. One way of telling if this is happening would be if the request latency varies wildly even though conditions aren't changing.
Again: I for one am implying that NT was tweaked for this test into a form that bears no resemblance to a web server- and Linux will never, ever catch up to it without similarly discarding any concern for the actual requests. Paying no attention to latency is just faster- checking on times costs, sorting costs, and serving pages upon requests totally bypasses the gains you get from caching and spitting out identical pages in quick succession. The trouble is, as I said, this course of optimization leads directly away from realworld performance and towards a thing that only works with benchmarks...
Does anybody, anywhere, have records on the statistical distribution of latency re. this NT optimization, compared to the tested Apache setup? Will they talk about latency at all, or try very hard to make no comment on it? Are there any records on _regular_ NT latency that might provide a clue? My experience trying to download from NT servers does suggest that latency varies wildly...
My experience in regard to NT is very similar. For the last 3 years working with various NT installations I had about 10-15 crashes and just about all of them were traced to bad drivers or hardware problems. Simply put NT server is quite stable. As for Linux, I had seen kernel panic (actually reboot ) once or twice and it was hardware problem too. So I guess both systems are quite stable. ..) there is no other choice but reboot. Linux is much more stable than Windows 95 but can't say the same when comparing Linux to NT. ...
Surprizingly, the biggest problems I had were related to MS GUI software (IE 4 etc.. now taking about NT workstation) What happens is that crash of certain MS software tends to make whole system unusable not because kernel itself went south but rather that many other user level services become unusable and essentially there is no other way to restore them other then reboot the machine. Theoretically it does not involve the kernel but practically the outcome it pretty much the same. To tell the truth Linux is not much better ( X is a joke as compared to Win GUI.)
Assuming one does not have any access to Linux box other than X, when it locks up ( and X does it much more often than NT workstation
I always looked at linux at cheaper and sometimes more convinient way to implement certain services.
It is a tool just like NT
I've got an idea. How about a fair real world test? An independent group comes up w/ some real world type need. They then give (or allow) each group to use some reasonable set amount of funds. That set amount will go to cover hardware, software, and personnel (they will have to pay themselves on set hourly figures that the independent group finds reasonable). That way you bring it down to the bottom line. The servers would be tested on how fast they could be and how reliable as well and of course price would also be tested since the cheaper system setup would allow more expenditure for greater overdesign.
Well, let's see, 99.9% uptime: out of 8766 hours every year, 99.9% uptime is 8757 hours, which means that the box is down 9 hours during a year. I remember some comments on a recent column linked by /. about the eBay thing which said that for _mission__critical_ apps, nothing short of 24/7 availability is acceptable. Now consider the many *nix boxes that (w/ competent administration) are up for over a year at a time. Consider a common situation: a customer calls, thier server is not working, you have to drive all the way across town (what at least an hour) and then reboot the box and wait for the filesystem to check itself, only then you can even start diagnosing the problem. (Which you usually won't find) End result: over two hours where an entire office has not been able to get any work done. Just imagine all the bitching and complaining. Who cares if i can *point my finger* at MS or HP or whoever, we've just waisted a lot of time and productivity and put everyone in a bad mood. I could care less about marketing and more about technical facts: a simple e-mail server running Linux _won't_ crash without some prodding, and then you have something more substantial to point your finger at.
Reid G. Ormseth, Esq.
Actually, this kind of touches on an interesting point. A while ago, someone (O'Riley?) was upset about Microsoft's licensing agreement for NT Server vs. Workstation (i.e. for workstation, you can only legally have 10 connections at a time, thereby making it useless as a commercial webserver). This was enforced by the system, and to counter MS' claims that it was for technical reasons, O'Riley (or whoever) found out how to change the system "type" (simply a registry entry) by disabling the 2 watcher threads that kept an eye on if you tried to change it yourself with regedit or some other means. They then proceeded to use the Workstation installation as a Server with no problems, and performance was probably equivalent.
When video-type operations moved to the kernel (as the AC above noted was a MAJOR change from NT3), it became a MAJOR reason why many people did NOT upgrade from NT 3.51, and why MS still supports it (since I guess they recognize the increased instability of NT 4).
IMO, what MS should've done (and may do in the future, I dunno, I've been out-of-touch with the MS "scene" for about 1.5 years now), is really made a technical difference between Server and Workstation. I.e. Server should retain slow video (it's a server for God's sake, why the hell does it need faster video) for increased stability, and leave faster video to Workstation. Makes perfect sense to me. And there's no reason why similar optimizations (performance vs. stability for everyday vs. server use) couldn't be done in almost any OS.
Have you sent a bug report to MS btw? If you can reproduce it, you should.
> 1. They claim that best hardware for Linux was chosen.
What they mean is that they we willing to replace a RAID interface for which Linux has no decent drivers with one from another manufacturer that has.
As others have pointed out, why not give each side a budget and let them chose the hardware for the task? There are other architectures than i386. There are other ways to use multiple processors than just SMP. There are other ways to get fast I/O other than 4 ethernet adapters in a single machine.
> 2. This configuration isn't 'exotic'. AFAIK It is _typical_ for big network server or web server ( not the one you use in your garage )
NO! It is typical of the configuration you absolutely want to avoid for a big server. It uses outdated processor architecture and a desktop oriented system architecture souped up to the max. It's expensive because everthing is special and bleeding edge, and there is no upgrade path because everything is maxed out already.
If you want a big server you either want to cluster a set of more modest servers (taking advantage of the low price and easy availablity of such systems) or get something from Sun, Dec, Sgi, or IBM that was designed from the start for much higher performance levels.
Just calling something 'enterprise class' doesn't make it a sensible choice.
> 3. NT IS faster than Linux, even on single-processor machines -> TCP/IP stack problem isn't the only one.
Still running 4 ethernet cards no doubt, still hugh quantities of memory, and still completely saturated with a huge number of short I/O requests. Yes, NT has a number of specific enhancements to handle this kind of load, most likely because NetWare kept beating the crap out it in high load file server benchmarks.
Choose somewhat different test parameters and Linux will be faster. The point is that subset of realistic configurations where Linux is faster is getting bigger and bigger. Proclaiming NT to be faster whereas in fact this is true only for a increasingly small subset of configurations running specific tasks is not exactly 'the whole truth'.
4. Mindcraft tests were biased.
They specifically selected a realm were NT was faster, then proudly announced they had found NT to be faster, and pocketed the Microsoft money. That's not bias, it's deceit.
Depends on what you mean by Linux. Linus thinks it's everything, including the software. And Linux "doesn't contain any unix code" is unlikely. Programmers would have studied Unix and inadvertantly used the same code/algorythms in the Linux kernel.
"Linux has a much lower per-process overhead than NT does, therefore making multithreading within the same process merely an alternative to multi-process multitasking" ;P
Bingo: and for what it's worth, MacOS has much worse per-process overhead and context switching time than either (I use it, and I never ask too much of it- works fine for running an application or two) It's good to see glimpses of reality through the onslaught of mindless technical hype. Because NT threading is a necessity and works pretty well, suddenly it's a requirement and nothing else can work- if NT was the one with really good multiprocess multitasking, and Linux was the one with mad threading, multithreading would be called an ugly hack and we'd be hearing about locks until our brains ran out our ears
Thanx Paranoid: good reality check.
Excuse me, but unless you're running apache from inetd (which is stupid for ANY amount of traffic), apache uses the Linux kernel to accept connections, and passes the connections to its child processes (subservers), which takes advantage of Linux's preemptive multitasking. This is exactly what you were claiming NT does that makes it better than linux. And if the question is one of multitasking vs. multithreading, I don't see why moving all the children into a single multithreaded process would add any performance gain, and would make the apache code much more complicated, because it would have to deal with mutexes and locks and all that un-fun stuff.
Uh, no. I wasn't saying that it made NT better than Linux...it was a response your your earlier argument that NT left some users idle. It was nothing to do with NT is better than Linux cause. It was to do with, uh, Linux isn't better than NT cause . And overheads for processes as opposed to threads is always much higher. Especially when you have to start communicating across processes. Where did you get your assumptions that NT handles processes worse than Linux?
Threading isn't just a nasty hack to get speed, it's there cause it's a good design - why do you think a mutithreaded ip stack is going to be incorporated into linux soon?
There is no necessity in NT to use threads, but obviously (as i outlined above) it would be faster. Certainly, in IIS, you can choose to have each session spawned off in a seperate process if you want to - threading just makes more sense. And I wasn't saying that Linux IS guilty, I was saying that it's more likely linux would be guilty of leaving requests idle than what you said about NT (eg. don't make stupid claims about NT).
You make the argument that while NT has it built into the kernel, linux has it as a separate program which makes NT run it inherently faster. Do you think businesses care about this? Unfortunately, I find it likely that they don't. They want results. Internals aren't important to them.
Overall though, I have to say that I'm fairly impressed with the attitude of the linux community this time around (at least so far). The general attitude seems to be that this test pointed out a couple performance problems in our favorite OS and that they'll now get tweaked. This is the kind of attitude that is going to help linux to gain acceptance.
Give a man a match, you keep him warm for an evening.
Light him on fire, he's warm for the rest of his life