A company called Stratus Technologies, that makes expensive, multiply redundant server hardware, guarantees that its high end four processor machine running Windows 2000 Advanced Server (cost: $150,000) will not have a hardware or OS-related failure or they will pay you $100,000. Stratus works on the device drivers to further ensure that they will not bring the OS down. See a Stratus Press Release for more information.
so the actual price is $50 000 then, right?
I don’t suppose it’s 100,000 per crash, is it? 🙂
the printf bug is counted? (not yet fixed)
No,
It’s for the first year – it crash only once and you will spend the rest of the year to bring it back.
printf bug doesnt crash my W2K SP3
so for anybody that need that machine (50k is worth the stuff anyway?)
just make a printf with some char a tab and loads of deletion and just cicle that for a 100 or more times just to be sure…
I did not fully understand what kind 4-way SMP was it.
It that 4-way Itanium2 or just regular xeon with three scsi cards for as little as $150,000?
Also, what the is failure? If I bring it down to crawl with memory leaks, is that a failure? I’d love a lawyer to see that article. Could I get $100K then refund for the original $150K? I would love to try then. Comes our roughly 66% for two weeks, that is including shipping and handling.
they realy want to charge 50k, but they came up with a cool marketing tool….give them 100k if the server fails…..sooooo the price becomes 150k.
Microsoft Windows can barely run itself.
Take Windows 2000 or Windows XP. Run as many instances of Internet Explorer as you can, loading some complex web page like espn.go.com. The machine will get slower and slower… eventually you will not be able to do anything else on the machine. It doesn’t matter how much RAM, how fast, etc.
Your machine will be fucked.
One can do the above trick many different ways. Run Office + IE, other apps, etc. With the design of Windows having many hard-coded resource limits, it is easy to crash a Windows box. Another example. On many Windows boxes, if you run out of disk space on your boot volume, the OS will think your disk is corrupted. Talk about stupid.
#p
“A $100,000 Bet on Windows”
Yeah, that’s a solid business model. I can’t wait to see it bluescreen in their face.
It’s like pointing a loaded gun at your shoes, pulling the trigger, and betting you won’t get shot in the foot. And you are wearing your shoes …
I can understand the hardware side. But to bet on a product that is not of your own control?
It’s like Ford betting that the gas in the pump is always going to be good for your engine …
It’s like betting …
It’s like betting …
haahaa they will not last for long
Now, why would somebody buy a 150k server and load IE with espn.com or printf ? Be serious for a moment. They can run sql server for example, that’s where SMP boxes are used and it won’t crash, with a good DBA. It works pretty well.
All operating systems can be crashed, but that’s not the point here.
Text from their press release:
>This server is approximately one-third less expensive than a >UNIX® cluster and one-fifth the cost of a Hewlett-Packard >NonStop™ server.
…HoHo – don’t compare ANY Windows driven/PC based server with HP’s (ak. Compaq’s, ak. Digital’s, ak. Tandem’s) NonStopâ„¢ server.
The architecture and scalability is so different, that it is a joke to compare these two patforms.
By the way: NonStopâ„¢ Systems are normaly used in environments where an outage of the system cost much more than their $100,000 offer. No NonStopâ„¢ customer will jump on that waggon. Sorry.
Ralf.
Microsoft Windows can barely run itself.
Take Windows 2000 or Windows XP. Run as many instances of “Internet Explorer as you can, loading some complex web page like espn.go.com. The machine will get slower and slower… eventually you will not be able to do anything else on the machine. It doesn’t matter how much RAM, how fast, etc.
Your machine will be fucked.
One can do the above trick many different ways. Run Office + IE, other apps, etc. With the design of Windows having many hard-coded resource limits, it is easy to crash a Windows box. Another example. On many Windows boxes, if you run out of disk space on your boot volume, the OS will think your disk is corrupted. Talk about stupid. ”
If you are running a web browser or an Office application on a server then you are an idiot.
Then again, if you are paying $150,000 for a machine and it is not running a Unix you are an idiot also.
Well, if they actually believe they have a reasonable MTBF calculated, all the more power to ’em. No different from insurance calculations. The cost of proving it’s not your ugly app that caused the mess will exceed $100,000.
There are a few words in small print next “money where mouth is”
1. Price is $180,000 and not $150,000
2. The CPUs are Intel Xeon MP with hyper-threading.
3. If you deliberately bring the server down, you void the guarantee.
4. There are no returns.
Basically, they charge $180,000 for about $25,000 worth of merchandise. Then when it breaks down they refund $100,000.
I best case scenario – you never even power machine on. They reap you off for $155,000 otherwise it may only be $55,000.
If anyone knows people interested in buying this crap, please post there enails to the osnews. I have brooklyn bridge laying around for only $50,000.
Hey I want to know about this crash ! Lu-zero do you have any example how to trigger that ? thanks !
> Take Windows 2000 or Windows XP. Run as many instances of
> “Internet Explorer as you can, loading some complex web page
> like espn.go.com. The machine will get slower and slower…
> eventually you will not be able to do anything else on the
> machine. It doesn’t matter how much RAM, how fast, etc.
My Linux box does that everytimes I forget that I don’t have to open too many mozilla windows.
Fabio Alemagna
aww jeez! the anti-MS anti-common sense trolls are here!
if you got a job working as a real professional instead of a perl-scripter running testcases, you wouldn’t troll about MS, Sun or Linux
The world of big scalable crunchers still belong to VMS, Tru64, IRIX and few others. People who require scalability and most important reliability will not even consider Windows servers.
//Take Windows 2000 or Windows XP. Run as many instances of “Internet Explorer as you can, loading some complex web page like espn.go.com. The machine will get slower and slower… eventually you will not be able to do anything else on the machine. It doesn’t matter how much RAM, how fast, etc.//
‘Scuse me junior, but that will happen with *_any_* desktop OS in the known universe.
If you say “not with *nix” you’re a liar as well as an idiot.
If you say “this article was talking about a server, not a desktop” … why the hell would you do that on a server, anyway?
Begone, fool!
IF the osnews parser does render correctly the escape…
<code>
while(1)
printf(“abcd ”);
</code>
I hate to say it, but Windows isn’t too much worse in reliability than most UNIX like OSs. Maybe just a little worse, but not a lot. Zelots will have to start picking other reasons to laugh at Windows soon.
Of course the $100,000 is a joke. Mr. Love probably has the best comment on why.
And Ralph was 100% dead on when he said nothing compares to HP’s (ak. Compaq’s, ak. Digital’s, ak. Tandem’s) NonStopâ„¢ server. That thing is not 99.99999999% uptime, it is 100.00000% uptime. It is what you run stock markets and phone companies and anything that cannot fail for even a little bit on. They are really amazing. For a fuzzy overview try this link:
http://nonstop.compaq.com/view.asp?PAGE=NSHimTech
Stratus is getting an average of %99.9998 uptime on all of their site installs. Take a look at their uptime meter here:
http://www.stratus.com/uptime/ftserver.htm
Stratus’ heavily customized hardware is on par with the Tandem (now Compaq/HP) NonStop architecture. Stratus is one of the most reputable fault-tolerant systems vendors around and they have Win2000 boxes running critical financial apps in Switzerland (where they are HQ’ed) with no problems. Also, they have several custom elements that they apply to the stock Windows 2000 install that makes it more reliable on their platform (among other things) because they undoubtedly have access to a good deal of the Windows source code that is relevant to their needs.
Hmm.. if 99.999% uptime equals 5.25 minutes of downtime in a year[0], they must have some super new security patches and service packs from Microsoft. You can’t tell me that a big machine like that boots in less than 3 minutes, at the rate of once a month for a patch, that’s 15 minutes of downtime/year.
[0] their numbers, not mine.
amen
When you are talking about Nonstop Sytems, you probably mean the old Tandem servers, which, for all intensive purposes, were VERY stable and VERY reliable.
Hmm, the good old days when quality meant something and one could hype their product unless there was something hype-a-ble.
” if you got a job working as a real professional instead of a perl-scripter running testcases, you wouldn’t troll about MS, Sun or Linux”
HAHAHAHAHA! I make more than you, your mom, *AND* your dad. Strange too because there’s lots of Perl involved.
(haha)
It is clear that 90% of the people posting have never worked a day in an IT field where high availability is a factor. It has become so easy to spew Non Microsoft pro Linux crap without the slightest bit of insight. In my company we have approximately 6 of these and to date all promises have been met. We also have HP-UX, Sun, and Linux Servers.
Stratus was making other highly available systems way before this plaform came up. In fact if you would read about the rest of there products they also support highly available systems running VOS and HP-UX so they are not really new to this whole concept. I am pretty sure the people who wrote this type or warranty are far more intelligent the the game jockeys that follow this site. I am sure if we needa a highly available Counter Strike Server I may ask this board.
The long and short of it is Stratus is delivering a fault tolerant hardware solution that support Win2k. My biggest complaint about these systems is the lack of flexibiliity. In my experience all 3rd party software that you wanted to install on the FT’s had to be approved by Stratus to prevent voiding the warranty. In addition, they say 4 processors but we benchmarked these against a dual processor Compaq 8500’s and the Stratus barely out performed the Compaq. There is alot of overhead on the systems to maintain this level of fault tolerance.
“Hmm.. if 99.999% uptime equals 5.25 minutes of downtime in a year[0], they must have some super new security patches and service packs from Microsoft. You can’t tell me that a big machine like that boots in less than 3 minutes, at the rate of once a month for a patch, that’s 15 minutes of downtime/year.”
Do you have problems understanding simple math, Matt?
The uptime meter that I posted the link to is a calculated AVERAGE of their entire installed base that are under warranty. It is NOT a number for a single machine, obviously. Why not learn to think before you open your mouth?
Guru or no guru. I design fault-resistant OSes for living.
There is very little need to systems like one being sold by Stratus for $180,000.
Yes, it is possible provided you have total control over software installed to ensure high uptime. However:
1. 100% is just a dream. All the do is play odds. Guarantee means nothing. I have had crappy Linux and *BSD boxes stay up for way over 365 days.
2. Buying Stratus stuff is hardly money spent wisely. Because performance (reliability) you get comes out at higher cost than for instance NonStop solution.
3. Institutions that need such systems – government, financial and phone companies (mostly) usually need performance along with availablity. Something Stratus cannot give you.
4. Only case when you would need such system is when some stupid IT manager decided that they need something be specifically designed for Windows, then you end up with having to buy whatever is available. In this scenario Stratus is acceptable, although overpriced solution.
Most of these anti-microsoft posts are simply crap. There is no insight, no reasonable logic, almost all of them are from people who don’t know enough about computers, operating systems.
If you open 100 Mozilla windows at the same time in Linux with complex pages, and if you still have Mozilla not crashed yet, what do you think that will happen to your OS. What does this have to do with OS itself. The OS does what you want. You can’t complain OS is less responsive when you run 100 jobs.
The Prophet, i realize it is a must for you to bash Microsoft – but this is a server. It wouldn’t run an desktop OS. It is running Win3k Advance Server. Which is very different.
Sergio and others defending Microsoft solutions,
We, people who work every day in big IT companies on *NIX, VMS clusters have completely different perception of reliability and stability. We don’t troll you.
Reliability must be 100%. Not 95%.
Daemons running on customer sites (mobile operators) must be 100%. One node down in the cluster ignites full investigation and numerous customers reports.
Our development environment is Tru64 and VMS on 64-bits multi-node alpha clusters. They’ve been not already rebooted for 2 years and are constantly “thrashed” by about 20 developers with all imaginable “attacks”.
After all, Microsoft will probably reach this level of reliability as well, while players like Tru64 dissapear in past…
aka, we don’t defend Microsoft. Microsoft doesn’t need our defending. Most people who you blame of defending Microsoft simply try to point out the truth.
If you make up stuff about Microsoft, or any other company or Linux the same people will come and will show that those statements are meaningless, pointless, wrong whatever.
The point is that, you are not the UNIX guy and we are the Windows guy. You even got that wrong. I myself run Linux, Solaris here, don’t like Microsoft much, like open source a lot.
I couldn’t also understand how you dispute my point. If you run lots of jobs, your OS will be less responsive, and this is true for almost all Operating Systems, unless the OS doesn’t let you do that, which is not the case for the operating systems we are talking about here. So what you are saying and what I am saying is totally different, and it seems that you simply get angry over my comment which seem to defend Microsoft.
I don’t care about Microsoft, just don’t troll or make up stuff.