The EE Times also has an article about next gen server technology IBM (via Sequent) and some info about Sun's next-gen stuff. As usual, Sun are saying very little. From what I've heard seperately though, Sun are working on both a form of NUMA and something else called COMA (Cache-Only Memory Architecture). They might be doing both (on the same machine) for their next-gen server - Project Serengheti, because NUMA is good for some types of applications, while COMA is good for others, so by doing both, the end-users can choose which memory architecture best suits their needs.
Btw NAS costs $35,000 PER CPU! They cite "lack of demand" as a reason for not doing the Linux port (yet). Btw they only sold 75 copies in the last 3 months. (and that was their best yet). Given the cost, it's not surprising Linux demand is low - it's an app for really high end stuff.
Here's a press release about the next version, to ship next month.
She's been coding most of the time since she was about 20. She gets paid quite a lot and deserves it too - she's very very good.
I've started programming a "mere" 15 years ago.
We used to live in South Africa, and in 1980 the company she worked for had a dedicated line built to our house (to connect a terminal to their mainframe), so she could work from home while raising us kids. First time this happened in the country I believe. That's how much they valued her. (such things are far cheaper these days of course...)
Bit of a shame in some ways... however, I expect they'll effectively be giving lots of their tech to the open-source community. If SGI are dropping their support for Irix, that wouldn't be very nice to their some of their existing customers, because that means they are forced to port/convert. Some of you might think I'm nuts for suggesting it, but it's true. One of the many reasons Sun have been doing well is that they're quite clearly NOT moving away from SPARC/Solaris.
According to some people I talked to, SGI doing the Intel/NT Visual Workstations hasn't helped with their image with established customers. Didn't help either that they had/are planning to move from their own CPUs to IA-64. (basically, this makes them 'just another' Intel box-shifter in some people's eyes). Incidentally, the current CEO of SGI is an ex-HP (co IA-64 developers with Intel) exec. You'd be surprised what difference image can make. I heard of a Alpha based server Compaq were selling. They then decided to sell the same machine under the Digital brand. Sales increased 10x or something.
Not only is there jini.org, Sun have setup jiro.com (there is a jiro.org too, but that bounces to jiro.com) for sorting out a platform neutral high-end storage management framework. Basically, it seems Sun's "community source" method of software development is becoming their preferred method of developing software.
They haven't outright and clear said so yet, but there are very strong hints that they plan to make most, and perhaps eventually, all of their software "community source".
At the JavaOne developer conference recently, this is what Bill Joy said:
Finally, in order to get all of this to work, we've been searching ever since we started Sun almost 20 years ago now for finding ways to work with you all. The JavaOne conference is an incredible example of a community of people getting together and doing great things. We originally started with the idea of "open systems", which means that we'd publish APIs and sell implementations, but that many people could produce implementations of those APIs.
Next there's this idea for the Java platform. We tried to create a community and protect it from a predatory company that we were aware of that was likely to try to attack us with contract law, and discovered that contracts sometimes aren't enough to protect us because not everyone thinks they apply to them.
So what we've decided to do going forward is to try to work from a notion of community. You've seen the Java Community Process (JCP). The JCP allows stakeholders in the different areas, like the people who care about realtime to define the realtime stuff, and that's a really good thing.
More recently we've done Community Source, which is an attempt to blend the best things about open source and proprietary models together with the added benefit from open source that when you take a Community Source license, you're allowed to make proprietary enhancements to it. We still insist that you leave the APIs open, but you can take large chunks of commercial money and make commercial investments. This works for companies.
The open source model works for other communities, and for them it's great. But we wanted to come up with a model that would work for traditional companies as well, so that we could quickly move into Community Source as much of our intellectual property as possible, and hopefully all of it going forward. But we insist also that people remain compatible.
So Community Source has an additional right and an additional responsibility relative to open source. We've done this with a lot of our technology already, including picoJava, and Java and Jini technologies. We'll be doing it with more.
I think they way they're going with Jini is pretty good. They could do better for Java though. They're being kinda closed about what they're going to open though, and have only dropped hints and not made definite statements on their website.
However, some things they have hinted/said are: they will open up Solaris later this year. Also, on one of their Solaris pages they say they'll be making a new version of Solaris (presumably Solaris 8) available under their "easy access" (ie beta) program - they've never done that before. They also seem to be working on making their C/C++ development software and compilers available, and to Linux users as well - to help develop code that works on Linux and Solaris more easily. They've also made other things not mentioned in that article available under their "community source".
ESR makes the argument that product cycles are not as long as they were. This is true, though it misses the point that important bits can stay the same between generations. Eg say you invent wizzo-bang method to help 3D graphics and no-one else figures it out, and it gives you a big lead, for your next gen cards, you might just use a more refined algorithm. By publishing the original specs, you could give your competitors a helping hand. Of course, it'd still take them another year or more to actually bring something new to market, by which time you've got your 3rd gen out, but if your competitors have got a lot more money for research than you, then they could catch up.
However, even this is not a real excuse - if you have a situation like this, publish as much as you can, and guard your most treasured secrets if you want to be that way.
For example, Matrox released the specs for their graphics G200 cards, except their triangle setup engine (for whatever reason) and they've also released most of the specs for their next gen G400, even before it started shipping!
Over at VNU Use rs slam anti-virus vendors' attack on Unix security. Basically, some people who sell anti-virus software made some interesting claims about virii on Unix. The article quotes 80,000 different virii for Windows (that few?), and 1 for Unix in the last 15 years.
Some people make arguments that the only reason for this is because Windows is more common/popular. Yeah... 80000x more common? I think not. Besides, there would be a certain presige in writing a Unix virus as they are so rare.
To my knowledge, Sun have not (yet) announced or commited to releasing the source to Solaris. Last I heard was just thinking, strongly, about it. Do you have any references?
It's hard to tell just how much influence others had, but Scott McNealy and other big-wigs are quoted as contributing. The document comes across as a serious proposal and not just one-man's foresight.
I don't really know what happened after this, but I guess Novell couldn't be pursuaded to go along - they sold their IP rights to SCO...
It's kinda weird how they mention Solaris when I don't think it really existed, as a product, at the time. Solaris, or more precisely, Solaris 2, was a re-write of Sun's OS, with some fundamental changes. Their main OS was Sun OS 4 at the time - Sun OS 5 is Solaris.
I'm not exactly sure how things played out after that document. I know that Novell sold the Unix stuff to SCO sometime around then - maybe SCO offered more than Sun? I do know that a few years back Sun spent something like $80M to buy permanent (or 20 years I or something) licenses for some Unix/System-V stuff. Kinda funny how things turn out.
There have been noises from Sun (mostly late last year, early this year) about opening up the source to Solaris. (in that, everyone can get it, not just developers with money, or people in education). They said the biggest problem is licensing issues - Sun don't own everything in Solaris. Looks like things haven't changed much... BTW Java, Jini, designs for the SPARC and Java chips, are not that only thing Sun have put under their "community source license" recently. Actually, it seems for quite a few products/tools, when they announced new versions they also annouced that it would be out under their "community source license" too. If/when Solaris does go this way, it'll probably be first with Solaris 8 (which is expected early 2000), and might not be back-ported.
This is a bit off-topic, but has anyone heard of drivers for DV capture cards? ie so you can get video digitally off your DV camcorder or whatever. I was looking the other day and didn't find anything at the time...
It's also not that good. (think of the recent NT vs Linux benchmarks where the webserver only servers static pages...)
However, the "new and improved (tm)" SPECweb99 is almost ready: See the SPECwww99 home page
Some thoughts/comparisons
on
1GHz Alphas
·
· Score: 5
Here's some current SPEC results Some high-end SPECint/SPECfp results: PA-RISC 8500 @ 440Mhz 34/51.4 21264 Alpha @ 500Mhz 27.3/57.7
Some notes: AMD's K7, even though it has a better FP unit than P-II will 'only' get SPECfp of about 20 at 600Mhz. (don't have published info, so making guess based on that it's about 30-40% faster than P-II). Sun's next gen chip (UltraSparc-III) will apparantly get SPECint/fp of 35+/60+ at 600Mhz (no actual results yet) - it is supposedly being publicly shown at the DAC (Design Automation Conference) now, but won't ship in volume until end of the year.
Anyone know how much the 8500 costs? It has 1.5Mb of level 1 cache - it has 150M transistors, to the Alpha 21264's 'mere' 15M. It must cost loads... This cache probably skews the SPEC results quite considerably when comparing to 'real world' cases - the SPEC marks scale pretty well with cache size I've heard... I also see that the 8500 doesn't seem to scale at all well at SPECfp as you add extra processors, compared to the other chips.
Real world usage can vary immensly from the SPEC values, depending on what you're doing. I have friends who've compared various machines for high end computations (fluid dynamics) and they found the SGIs ran/scaled the best, even though they didn't have the best SPECfp results for a single chip - it's their massive data buses that do the trick. Actually, the PA-RISC 8500 doesn't have a complete Fortran compiler yet... Most people I know consider the Alpha to be let down by it's IO/bus data-rate,etc. Yes, it's better than PC, but it's not much compared to the other high-end RISC guys, especially SGI, though I expect this difference to change...
A final note, a problem you get with high-speed processors is that they become nice little microwave transmittors (at 700-800+ Mhz, I think it was) and so you really need to reduce the power (the PA-RISC 8500 consumes 85W @ 440Mhz) and up the shielding when clocking at this rate, because otherwise you'd get a REAL pizza cooker/toaster in your computer...
Make sure to read the link about why eBay probably preventable - it seems eBay didn't suffer from load problems, just that eBay hadn't installed an easily available patch - apparantly it'd been available for about a year. This LA Times story says about the same thing, but is much shorter.
The first link basically says that the eBay guys weren't paranoid enough about making sure the setup was reliable. This is always a problem. (hey, I'm working on a commercial web site that only got a proper sys-admin 2 years after it started...). Little side-note - one guy says Sun's clustering stuff is not that great... I know Sun have been a bit late in starting doing clustering stuff, but I've also heard that what they have done is pretty good, *shrug*. Actually, they just annouced version 3 last week, which also allows clustering of 16 Starfires, for 1024 processors. (they're also making the source code for this available...)
I haven't done much benchmarking myself, but I know some people who have, on serious code. The simple answer is that it's pretty decent for x86. gcc was pretty poor on RISC architectures (both much slower to compile the code, and the resulting code was less efficient too) though egcs is much better on the RISC side. Most admin/guru type people I know consider gcc to be a not particularly efficient compiler, but very cool for being damn portable.
I know someone who did some benchmarking with various compilers on high end RISC hardware (SGI Origin 2000 and Sun Starfire). It was doing fluid dynamic modelling, and the code was about 400,000 lines. The egcs compiler was about 3x slower at compiling the code. For actual generated code, for a variety of tests/setups, the egcs code was much slower on several tests (9x slower on one test), and on one or two tests it was actually slightly faster than the commercial compilers. On a few tests the egcs code failed to work due to bugs in the egcs compiler.
Compiling for high end hardware is very hard, and you generally need compilers that know the hardware to get the best results, which is why it's not that surprising egcs/gcc didn't do too well on the high-end hardware - because it rarely gets used for such things.
Sun have a long, detailed white paper on egcs VS Sun's compiler. They quote 34% faster SPECint code and 127% SPECfp code with their own compiler. They also promote some other things - better development environemnt, hence improving productivity. btw, Sun's standard compiler costs $500 and their pro one cost $1500, I believe... (of course, if you're paying developers to write code, saving two-man weeks of time could be enough to justify that $1500, for the guy working from home, the cost is prohibative)
Sun have said they're working on some stuff to help people writing software that works correctly on Solaris and Linux more easily. It'd be nice if they made their compilers work under Linux and be free to non-commercial useage.
I don't really like writing flames, but that article was just sad. Even attempting to debunk it would be a waste of time. We don't need this sort of trash, just pull the article please admin, and give us something interesting to talk about.
Firstly, let me say that if you're enterprise is running on Windows, then the cost of the OS is still going to be significant. If you're using Windows NT client, you're going to be paying through the nose. Still, compared to say, the cost of all the sys-admin involved with Windows NT, plus the hardware, infra-structure, training etc, it might not be that significant. Though, after the initial payment on the hardware (once off), the ongoing software costs are going to be pretty significant, as it's yearly. If you want to include hidden costs though (ie loss of work due to system outages), then we could be much crueler to Windows...
On the other hand, if we compare to proprietary Unix OSs, then the software costs are much lower. Unix software is generally considered expensive - this is often true for a single user, or small number of users, but often it becomes very cheap (per user) for an enterprise setup. For example, the RRP for Solaris workstation is about $400, though you can get it much cheaper than this - free for non-commercial usage. That's obviously more than Windows. However, getting an infinite user license for Solaris costs $1500 or something, meaning if you have 1000 users in your company, it costs $1.50 for each one for the OS. NT workstation is $250 (or something like that) per user, and you don't really get volume discounts.
Putting it on a different level... for the high-end/datacenter level, the hardware could easily cost several million. Compared to this, a couple of hundred for the cost of the OS is nothing. Besides which, the cost of the proprietary OS will be included in the cost of the hardware, effectively for free. (I believe you can get Sun hardware without the OS, which does give you a small discount - the cost of the CDs, books and other media, which is about $20 or something)
On another level... the cost of migrating can be pretty huge - first, you need to re-train your sys-admin (or chuck out the old ones, and get in some new) which can be pretty painful. Ditto if you have software developers. Then you need to re-train your staff, port your specialist software (if applies), etc etc. You also have to figure in the cost of downtime while this is all going on. This is why people are still using 'ancient' mainframes - because a) the risk of moving is too much, and b) it probably does what you need already anyway.
Unless I missed something, the IBM guy did not say that they were going to support Linux only now, just that it would be cheaper. This is pretty obvious - it's cheaper to support one OS than lots.
I think they would have difficulty trying to get everyone to use just Linux, in companies and deperatments where they currently have no Linux installations, and hence so sys-admin. This is for enterprise level stuff, so this is important.
At the moment, all they seem definate about is adding Linux support, to their current list. Migrating to Linux-only seems more like wishful thinking in the short/medium term.
(gripe mode on) If you're going to post stories like this, why couldn't you post one I submitted the other day - Sun licensing their JavaServer Pages/servlet source code to the Apache Group, and letting Apache distribute that source code under the Apache license. see the Jakarta home page and this JavaWorld article(grip mode off)
I don't really know very well what AOL get up to, but as it says in the subject, having some similar points is not the same as being the same. Similarly, having a monopoloy/large share of a particular market does not make one a monopolist - ie use that share to abuse consumers and other companies.
Does any of the other 3:
Keep changing the APIs between released products to make it harder for other developers to keep up with MS's own apps?
Using licensing agreements/pricing as a way to crush competitors, and encourage/force them to not compete with MS? (see how much arm-twisting MS did to IBM over Windows 95...)
Pay employee's little salary, recruiting only/mostly 'believes' from college with little real-world experience, using share-options as a carrot to encourage employee's to slave away mindlessly, and basically encourage the execs to do whatever they can do smash the competition?
Have Mindcraft analogue?
Have a history of partnering/investing in other companies, 'stealing' that company's ideas/etc, and bringing out their own product, in the end killing their 'partner'?
Have $20B in the bank ready to spend on crushing/buying up the competition.
I could go on much further. Basically, there's lots of things MS have done that just about no other company has ever done or done to that extreme. There's lots of people who seem to fall to MS's own PR - ie anyone who wants to supplant us, is just as bad as us.
Also, to focus on Sun's Java software is highly missleading. Sure it's their biggest PR problems with respect to the open source community, but it's still a far cry from typical MS operations. Anyone can get the source code to Java. (not even close with Windows, and MS's 'hints' of open source seem to be just starshine so far). Sun's 'community source liscense' is still pretty closed, and they need to sort out things though. Neither is Java a once off - Sun are making most of their source code available it seems. Jini is most definately not vapourware - how can it be vapourware if the source code has been available for downloading for months? Also, part of Java (the JavaServer Pages/servlet stuff) is be coming with Apache, as source code, under the Apache liscense. Far cry from MS.
Java is just a tiny, but highly visible part of Sun, and they make very little revenue or profit from it directly. Sun's hardware and Solaris software is highly reliable, secure, scalable - far more so than Windows.
If I remember correctly, isn't MSNBC owned/partly owned by MS? Anyway, the article certainly seems over-hyped to me. What has come up in the trail with regards to the Netscape/AOL/Sun deal hasn't impressed the judge certainly - he made that pretty clear. I've heard that the 3 are working on a consumer box, but it's still early days yet (ie might not even see it this year) and I really doubt it's aiming to be the equivilant or a substantial replacement of Windows - it's going for the lower-end market - I would guess they're going for a NC/Javastation type thing.
Unless someone has something that comprehensivly attacks MS in all the major sectors (consumer, business desktop, server), you're not going to hurt MS that much. That hardest is probably the business desktop.
A spokesman for Sun Microsystems said a problem with the tech firm's software contributed to the problem and that a service team was working with EBay to restore service.
"We know what the probem is and we're fixing it," said Doug VanAman, director of public relations. "We're investigating it further."
He said the glitch appears to be unique to EBay and not with any other Sun customers.
At the bottom it says "The site outage was unrelated to the Worm.Explore.Zip computer bug, which began infecting systems across the country this week." Excuse me why I die laughing...
Slashdot Mirror
The EE Times also has an article about next gen server technology IBM (via Sequent) and some info about Sun's next-gen stuff. As usual, Sun are saying very little. From what I've heard seperately though, Sun are working on both a form of NUMA and something else called COMA (Cache-Only Memory Architecture). They might be doing both (on the same machine) for their next-gen server - Project Serengheti, because NUMA is good for some types of applications, while COMA is good for others, so by doing both, the end-users can choose which memory architecture best suits their needs.
Btw NAS costs $35,000 PER CPU! They cite "lack of demand" as a reason for not doing the Linux port (yet). Btw they only sold 75 copies in the last 3 months. (and that was their best yet). Given the cost, it's not surprising Linux demand is low - it's an app for really high end stuff.
Here's a press release about the next version, to ship next month.
I've started programming a "mere" 15 years ago.
We used to live in South Africa, and in 1980 the company she worked for had a dedicated line built to our house (to connect a terminal to their mainframe), so she could work from home while raising us kids. First time this happened in the country I believe. That's how much they valued her. (such things are far cheaper these days of course...)
Bit of a shame in some ways... however, I expect they'll effectively be giving lots of their tech to the open-source community. If SGI are dropping their support for Irix, that wouldn't be very nice to their some of their existing customers, because that means they are forced to port/convert. Some of you might think I'm nuts for suggesting it, but it's true. One of the many reasons Sun have been doing well is that they're quite clearly NOT moving away from SPARC/Solaris.
According to some people I talked to, SGI doing the Intel/NT Visual Workstations hasn't helped with their image with established customers. Didn't help either that they had/are planning to move from their own CPUs to IA-64. (basically, this makes them 'just another' Intel box-shifter in some people's eyes). Incidentally, the current CEO of SGI is an ex-HP (co IA-64 developers with Intel) exec. You'd be surprised what difference image can make. I heard of a Alpha based server Compaq were selling. They then decided to sell the same machine under the Digital brand. Sales increased 10x or something.
They haven't outright and clear said so yet, but there are very strong hints that they plan to make most, and perhaps eventually, all of their software "community source".
At the JavaOne developer conference recently, this is what Bill Joy said:
Next there's this idea for the Java platform. We tried to create a community and protect it from a predatory company that we were aware of that was likely to try to attack us with contract law, and discovered that contracts sometimes aren't enough to protect us because not everyone thinks they apply to them.
So what we've decided to do going forward is to try to work from a notion of community. You've seen the Java Community Process (JCP). The JCP allows stakeholders in the different areas, like the people who care about realtime to define the realtime stuff, and that's a really good thing.
More recently we've done Community Source, which is an attempt to blend the best things about open source and proprietary models together with the added benefit from open source that when you take a Community Source license, you're allowed to make proprietary enhancements to it. We still insist that you leave the APIs open, but you can take large chunks of commercial money and make commercial investments. This works for companies.
The open source model works for other communities, and for them it's great. But we wanted to come up with a model that would work for traditional companies as well, so that we could quickly move into Community Source as much of our intellectual property as possible, and hopefully all of it going forward. But we insist also that people remain compatible.
So Community Source has an additional right and an additional responsibility relative to open source. We've done this with a lot of our technology already, including picoJava, and Java and Jini technologies. We'll be doing it with more.
I think they way they're going with Jini is pretty good. They could do better for Java though. They're being kinda closed about what they're going to open though, and have only dropped hints and not made definite statements on their website.
However, some things they have hinted/said are: they will open up Solaris later this year. Also, on one of their Solaris pages they say they'll be making a new version of Solaris (presumably Solaris 8) available under their "easy access" (ie beta) program - they've never done that before. They also seem to be working on making their C/C++ development software and compilers available, and to Linux users as well - to help develop code that works on Linux and Solaris more easily. They've also made other things not mentioned in that article available under their "community source".
They do seem pretty serious about it.
However, even this is not a real excuse - if you have a situation like this, publish as much as you can, and guard your most treasured secrets if you want to be that way.
For example, Matrox released the specs for their graphics G200 cards, except their triangle setup engine (for whatever reason) and they've also released most of the specs for their next gen G400, even before it started shipping!
It's going a bit slower than normal... but not /.'ed. This is from the UK too...
Sun sponsor the McLaren F1 racing team, and also provide hardware. see here. However, it doesn't use Java in the car, as far as I know.
Some people make arguments that the only reason for this is because Windows is more common/popular. Yeah... 80000x more common? I think not. Besides, there would be a certain presige in writing a Unix virus as they are so rare.
To my knowledge, Sun have not (yet) announced or commited to releasing the source to Solaris. Last I heard was just thinking, strongly, about it. Do you have any references?
I don't really know what happened after this, but I guess Novell couldn't be pursuaded to go along - they sold their IP rights to SCO...
I'm not exactly sure how things played out after that document. I know that Novell sold the Unix stuff to SCO sometime around then - maybe SCO offered more than Sun? I do know that a few years back Sun spent something like $80M to buy permanent (or 20 years I or something) licenses for some Unix/System-V stuff. Kinda funny how things turn out.
There have been noises from Sun (mostly late last year, early this year) about opening up the source to Solaris. (in that, everyone can get it, not just developers with money, or people in education). They said the biggest problem is licensing issues - Sun don't own everything in Solaris. Looks like things haven't changed much... BTW Java, Jini, designs for the SPARC and Java chips, are not that only thing Sun have put under their "community source license" recently. Actually, it seems for quite a few products/tools, when they announced new versions they also annouced that it would be out under their "community source license" too. If/when Solaris does go this way, it'll probably be first with Solaris 8 (which is expected early 2000), and might not be back-ported.
This is a bit off-topic, but has anyone heard of drivers for DV capture cards? ie so you can get video digitally off your DV camcorder or whatever. I was looking the other day and didn't find anything at the time...
However, the "new and improved (tm)" SPECweb99 is almost ready: See the SPECwww99 home page
Some high-end SPECint/SPECfp results:
PA-RISC 8500 @ 440Mhz 34/51.4
21264 Alpha @ 500Mhz 27.3/57.7
Some notes: AMD's K7, even though it has a better FP unit than P-II will 'only' get SPECfp of about 20 at 600Mhz. (don't have published info, so making guess based on that it's about 30-40% faster than P-II). Sun's next gen chip (UltraSparc-III) will apparantly get SPECint/fp of 35+/60+ at 600Mhz (no actual results yet) - it is supposedly being publicly shown at the DAC (Design Automation Conference) now, but won't ship in volume until end of the year.
Anyone know how much the 8500 costs? It has 1.5Mb of level 1 cache - it has 150M transistors, to the Alpha 21264's 'mere' 15M. It must cost loads... This cache probably skews the SPEC results quite considerably when comparing to 'real world' cases - the SPEC marks scale pretty well with cache size I've heard... I also see that the 8500 doesn't seem to scale at all well at SPECfp as you add extra processors, compared to the other chips.
Real world usage can vary immensly from the SPEC values, depending on what you're doing. I have friends who've compared various machines for high end computations (fluid dynamics) and they found the SGIs ran/scaled the best, even though they didn't have the best SPECfp results for a single chip - it's their massive data buses that do the trick. Actually, the PA-RISC 8500 doesn't have a complete Fortran compiler yet... Most people I know consider the Alpha to be let down by it's IO/bus data-rate,etc. Yes, it's better than PC, but it's not much compared to the other high-end RISC guys, especially SGI, though I expect this difference to change...
A final note, a problem you get with high-speed processors is that they become nice little microwave transmittors (at 700-800+ Mhz, I think it was) and so you really need to reduce the power (the PA-RISC 8500 consumes 85W @ 440Mhz) and up the shielding when clocking at this rate, because otherwise you'd get a REAL pizza cooker/toaster in your computer...
That first link should have read 'eBay problems probably preventable'.
The first link basically says that the eBay guys weren't paranoid enough about making sure the setup was reliable. This is always a problem. (hey, I'm working on a commercial web site that only got a proper sys-admin 2 years after it started...). Little side-note - one guy says Sun's clustering stuff is not that great... I know Sun have been a bit late in starting doing clustering stuff, but I've also heard that what they have done is pretty good, *shrug*. Actually, they just annouced version 3 last week, which also allows clustering of 16 Starfires, for 1024 processors. (they're also making the source code for this available...)
I know someone who did some benchmarking with various compilers on high end RISC hardware (SGI Origin 2000 and Sun Starfire). It was doing fluid dynamic modelling, and the code was about 400,000 lines. The egcs compiler was about 3x slower at compiling the code. For actual generated code, for a variety of tests/setups, the egcs code was much slower on several tests (9x slower on one test), and on one or two tests it was actually slightly faster than the commercial compilers. On a few tests the egcs code failed to work due to bugs in the egcs compiler.
Compiling for high end hardware is very hard, and you generally need compilers that know the hardware to get the best results, which is why it's not that surprising egcs/gcc didn't do too well on the high-end hardware - because it rarely gets used for such things.
Sun have a long, detailed white paper on egcs VS Sun's compiler. They quote 34% faster SPECint code and 127% SPECfp code with their own compiler. They also promote some other things - better development environemnt, hence improving productivity. btw, Sun's standard compiler costs $500 and their pro one cost $1500, I believe... (of course, if you're paying developers to write code, saving two-man weeks of time could be enough to justify that $1500, for the guy working from home, the cost is prohibative)
Sun have said they're working on some stuff to help people writing software that works correctly on Solaris and Linux more easily. It'd be nice if they made their compilers work under Linux and be free to non-commercial useage.
I don't really like writing flames, but that article was just sad. Even attempting to debunk it would be a waste of time. We don't need this sort of trash, just pull the article please admin, and give us something interesting to talk about.
see this image. It's from Quake 3, with all the effects on. btw it's 615Kb JPG...
On the other hand, if we compare to proprietary Unix OSs, then the software costs are much lower. Unix software is generally considered expensive - this is often true for a single user, or small number of users, but often it becomes very cheap (per user) for an enterprise setup. For example, the RRP for Solaris workstation is about $400, though you can get it much cheaper than this - free for non-commercial usage. That's obviously more than Windows. However, getting an infinite user license for Solaris costs $1500 or something, meaning if you have 1000 users in your company, it costs $1.50 for each one for the OS. NT workstation is $250 (or something like that) per user, and you don't really get volume discounts.
Putting it on a different level... for the high-end/datacenter level, the hardware could easily cost several million. Compared to this, a couple of hundred for the cost of the OS is nothing. Besides which, the cost of the proprietary OS will be included in the cost of the hardware, effectively for free. (I believe you can get Sun hardware without the OS, which does give you a small discount - the cost of the CDs, books and other media, which is about $20 or something)
On another level... the cost of migrating can be pretty huge - first, you need to re-train your sys-admin (or chuck out the old ones, and get in some new) which can be pretty painful. Ditto if you have software developers. Then you need to re-train your staff, port your specialist software (if applies), etc etc. You also have to figure in the cost of downtime while this is all going on. This is why people are still using 'ancient' mainframes - because a) the risk of moving is too much, and b) it probably does what you need already anyway.
I think they would have difficulty trying to get everyone to use just Linux, in companies and deperatments where they currently have no Linux installations, and hence so sys-admin. This is for enterprise level stuff, so this is important.
At the moment, all they seem definate about is adding Linux support, to their current list. Migrating to Linux-only seems more like wishful thinking in the short/medium term.
(gripe mode on) If you're going to post stories like this, why couldn't you post one I submitted the other day - Sun licensing their JavaServer Pages/servlet source code to the Apache Group, and letting Apache distribute that source code under the Apache license. see the Jakarta home page and this JavaWorld article(grip mode off)
Does any of the other 3:
I could go on much further. Basically, there's lots of things MS have done that just about no other company has ever done or done to that extreme. There's lots of people who seem to fall to MS's own PR - ie anyone who wants to supplant us, is just as bad as us.
Also, to focus on Sun's Java software is highly missleading. Sure it's their biggest PR problems with respect to the open source community, but it's still a far cry from typical MS operations. Anyone can get the source code to Java. (not even close with Windows, and MS's 'hints' of open source seem to be just starshine so far). Sun's 'community source liscense' is still pretty closed, and they need to sort out things though. Neither is Java a once off - Sun are making most of their source code available it seems. Jini is most definately not vapourware - how can it be vapourware if the source code has been available for downloading for months? Also, part of Java (the JavaServer Pages/servlet stuff) is be coming with Apache, as source code, under the Apache liscense. Far cry from MS.
Java is just a tiny, but highly visible part of Sun, and they make very little revenue or profit from it directly. Sun's hardware and Solaris software is highly reliable, secure, scalable - far more so than Windows.
Unless someone has something that comprehensivly attacks MS in all the major sectors (consumer, business desktop, server), you're not going to hurt MS that much. That hardest is probably the business desktop.
"We know what the probem is and we're fixing it," said Doug VanAman, director of public relations. "We're investigating it further."
He said the glitch appears to be unique to EBay and not with any other Sun customers.
At the bottom it says "The site outage was unrelated to the Worm.Explore.Zip computer bug, which began infecting systems across the country this week." Excuse me why I die laughing...