Sometimes, Google’s search engine does a better job of telling us about IT vendors than the vendors’ own public relations and marketing machines, which are often there mostly to deflect questions rather than answer them. So it is with the next commercial and development iterations of Oracle’s Solaris Unix operating system.
I expected Oracle to open source TCK and finally let Apache in (yes I am an optimist), also I am a big fan of NetBeans and I am kinda unsure how they will handle 3 IDEs, plus VirtualBox and glassfish, which is a decent container. But I wish that IBM had snatched Sun instead of Oracle…
I do sound like a broken record, cos I already said this here, but IMO Oracle is a bad influence in many of the technologies that I use to be paid every month.
I think that in the end I will bite the bullet and learn C#
You make absolutely no sense. First you start fear-mongering that Oracle may not be as open as you like, so your solution is to switch to Microsoft? Do you have any idea of what you are talking about? With .NET you are for all intents and purposes tied to one OS (in fact *the newest version* of that OS if you want the latest IIS), one IDE, one provider of middleware, so you are entirely at the mercy of one company. The Java stack on the other hand provides everything you need and might need in the future, compatible with any major OS and most of it *for free*? What on earth are you smoking?
Edited 2010-06-22 19:12 UTC
I have poor english skills my bad…
but back to the point, i have actually idea of what i am talking about. i was involved on two systems that were oracle top-down. everything running on solaris. and i had also to deal with their practices of “bullying” customers to upgrade their software. so i guess when oracle/ibm push customers around if okay but if ms does it is bad right? i know your type.
as much as i like java and related tools i want something with stability and clear development path. i dont see oracle changing the way sun behaved with jcp.i dont see how oracle will handle those tools. what i see is a clear divide in the “community” and no idea where to go next.
so you can cut all the rubbish “oh you dont have any idea of what you say”. in my opion you probably live of a freetard fantasy… i need to pay my bills and if i need to learn .net for that i dont understand what is your problem.
i guess you will be one of those cobol dudes…
ps. sorry for typos but i was typing this on my mobile.
According to Tiobe, Java is much greater than .NET.
I see lot of .NET and C# in my line of work. But that is only on the clients. All big Enterprise servers are developed in Java/C/C++, never C#. The reason is simple: a good implementation of C# exists only on Windows. And you never use Windows on big Enterprise servers. London Stock Exchange tried Windows recently, but is now migrating to Linux + Solaris because of problems.
You must use big Enterprise OSes on the large servers, different flavours of Unix or OpenVMS. If C# would have existed on these big Enterprise systems, then the server software would have been written in C#. But now that does not happen. On the servers we never see C#. C# is only good for clients.
-Big servers on Unix (with long lifespan) = Java/C/C++.
-Small client software on Windows (with short lifespan) = Java/C/C++/C#
What about MySpace.com? Is that large enough?
Please, I do not talk about a web site. London Stock Exchange could not keep uptime with Windows. And look at msn, it is down all the time.
How do you know the fault was with Windows and not within the software stack?
How do web hosts guarantee 99.9% uptime with Windows Server if it is unreliable?
Gmail had a lot of downtime last year, so I suppose the logical conclusion is to blame Linux.
This link negates your original claim
http://highscalability.com/myspace-architecture
Edited 2010-06-23 19:43 UTC
How do I know that the fault was Windows for LSE – London Stock Exchange? I doubt the software stack is as buggy as Windows. Windows has a bad reputation for being crash prone. C# has not the same reputation.
Why do you think that LSE migrates away from Windows? LSE spent 50 million USD to develop the Windows system, and only two years later, migrates to Linux and Solaris. Do you really think that someone throws away an investment of 50 million USD lightly? Havent you read all the articles about all problems LSE had?
Regarding MySpace, please. I dont talk about web sites. You can basically throw in another server and you get better uptime for a web site. The structure is easy. MySpace uses over 6,000 servers. A stock exchange has much higher demands and is much more difficult to maintain high uptime. And besides, 99.9% is bad for a stock exchange. Wouldnt do.
A stock exchange is going be spread over multiple servers. Maybe you should quit digging your hole even deeper.
http://www.embeddedstar.com/press/content/2003/2/embedded7349.html
99.9% is what web hosts guarantee, not an average.
Anyways this was your original statement:
On the servers we never see C#.
That’s clearly not true, and a stock exchange is going be spread over multiple servers.
http://www.embeddedstar.com/press/content/2003/2/embedded7349.html
What is your overall message here? You can use Windows Server for websites like Bing and and MySpace but nothing bigger? You have to use Linux for stock exchanges? That again negates your original claim of C# not being used on servers.
As for wasted money the Munich Linux project has been a a failure so by your logic we should blame Linux.
http://fixunix.com/linux/363315-its-official-munich-linux-migration…
You really should have stuck with “java is free” or something along those lines instead of trying to go with the 2001 talking point of Windows Server not being good enough for enterprise use. There’s countless cases that prove you wrong, and I’m not even sure why you keep replying to me when it is clear you were just pushing BS and were assuming that you wouldn’t be called on it.
You dont understand.
Yes of course a stock exchange are spread over multiple servers. Maybe 5-10 servers.
My point is, to run a website is simple. Easily scalable. Just add another server. It is like Linux used in supercomputers (which is basically a bunch of computers on a fast network) – Linux scales well in that case. The structure of a supercomputer is easy, just add another node. It is a totally different thing to run a single OS image on multiple CPUs – that is called scaling vertically. Scaling horizontally is trivial. But some workloads do not admit horizontal scaling – and that is very tricky to do well. Linux scales good horizontally (in a network), but bad vertically (on one single computer with lots of CPUs). Linux running on supercomputers does not prove that Linux scales well vertically. It only proves that Linux is good at horizontal scaling, which it is good at.
As I showed with links, AIX does not scale well. It is now being rewritten. AIX is an old Enterprise Unix and it does not scale well – vertically even today with lots of development. The difficult thing to do – is to scale vertically. To add nodes is trivial. Even AIX has problems with vertical scaling.
And guess what? Websites are scaling horizontally. MySpace proves nothing about vertical scaling. Now, do you finally understand that MySpace is not relevant? Please talk about more demanding stuff. Any fool can add a node to a web site.
And that sucks for really critical applications. Wont do.
Why is it not true? You havent disproved this. Can you explain again?
You linked to an antique old fashioned, bad performing stock exchange that migrated to Solaris almost ten years ago. And that Solaris exchange has really really bad performance, the old one must be much worse. Jesus. So what is your point?
My message here, is that you can use any shit OS you want, to do easy stuff. Just add another node and that will (hopefully) increase your uptime.
But some work loads need vertical scaling, and then you need to go to a few big servers. It does not help to add just a new node. Then you need an Enterprise OS, such as Unix or OpenVMS. And frankly, Linux does not really belong here. The Linux stock systems are run with low utilization, not stressed heavily. Linux is too fragile.
How does it negate my claim? A good implementation of C# does not exist on big Enterprise Servers. Therefore you need to run C# on windows. And Windows does not do when running difficult server systems.
I didnt get that. Can you explain your line of thought again? Munich does not migrate to Linux, and that means that Linux is not fit for large Enterprise systems? Or? Actually, I didnt get your point. What are you trying to say?
Let me say this again. Windows does not cut it in large demanding systems like stock exchanges. Why do you think London throws away 50 million USD and several years of development after only two years of running problems? No stock exhcange except LSE runs Windows. And not even LSE soon. The manager that chose Windows system is now fired. The new manager chose Linux and Solaris – like everyone else. To try Windows as a stock exchange was a very expensive experiment. Dont do that again. Windows is good for running web sites, and other tasks where you just add another node. Easily parallellizable tasks. Some problems are not parallellizable, though. Then you need heavy gear and a heavy OS.
Whenever someone says “Linux is too fragile”, yet provides no evidence, anecdotal or not, to back up that claim, I think of the movie Idiocracy.
“But… it’s got electrolytes…”
Oh, there are lots of links, for instance Linux is fragile and buggy, whereas Solaris is robust on the same piece of hardware. Linux sucks for large file storage work. Linux scales bad, Solaris scales better. etc etc
You wanna see some links?
I await with baited breath.
The productivity gains from ASP.NET make up for the licensing cost of Windows Server and then some.
Tying yourself to Java means tying yourself to a company that really just wants to sell database software and could care less about software development. Who knows how much they will let Java stagnate along with the rest of Sun.
Going from Java to C# is easy if you already know C. You can always go back if you don’t like it.
Edited 2010-06-23 18:21 UTC
Actually, that’s exactly what it *doesn’t* mean. Oracle is not the only provider of Java technology, even if they were to halt all further development of every Java-based product they put out (which of course they won’t), you’d still have an enormous, free, multi-vendor ecosystem at your disposal. I understand that for someone whose main experience is within the confines of MS’s walled garden, this may be hard to comprehend.
–double post–
Edited 2010-06-22 19:09 UTC
Opensolaris has mono. There is also pnet by gnu. Mono has a binding for firebird. I do not know if monodevelop is available for Opensolaris. However gnu emacs can help you. If you want to be a C# devotee you can continue on Windows. But I do not recommend Visual Studio. Finally if you want to keep Solaris on your partition take a look at the wonderful packages by activestate in order to learn some real programming.
I don’t like Oracle’s tactics either, but they do embrace open source software in their own kind of Ellisonian way. At least they’re not killing off Solaris, OpenSolaris, MySQL or VirtualBox. If it’d been Microsoft buying Sun, they would have killed all of them instantly.
I do agree that IBM would have been the better buyer.
As for the C# bit… ???
No, they’re just not saying anything about OpenSolaris, and not releasing any development builds, effectively cutting it off, it seems. They’re not killing it, they’re just letting it die of neglect. And, I’m not sure there’s enough resources to fork it, considering most of the development came from within Sun.
At least the BSDs are taking the good bits from OpenSolaris…
Have you ever been at a company right after it was acquired or after some big reorg has happened? Productivity effectively dies for months or years while the politics of the whole thing is figured out.
Just give it some time. I am sure they are not going to let OpenSolaris die, they just haven’t had any time to do real work on it for the past months….
Edited 2010-06-23 06:57 UTC
They already have another open source OS to support, Unbreakable Linux. OpenSolaris really doesn’t fit into their plans.
Anyone using OpenSolaris should move to FreeBSD. Oracle doesn’t care about OpenSolaris and would have killed it by now if it wasn’t for the bad p.r. it would elicit.
Edited 2010-06-23 18:40 UTC
Larry Ellison has officially praised Solaris and said that Linux is for the low end, but Solaris is for the high end.
Oracle database is developed on Solaris, and then ported to other OSes. Solaris is the most common platform OS for Oracle database.
I dont get it, how FreeBSD fanboys can hope that Solaris gets killed. From which OS will FreeBSD get new cool tech then? I would like to see FreeBSD developers come up with something like ZFS or DTrace. Both ported to FreeBSD. And instead of being happy and thanking OpenSolaris team, FreeBSD fanbois hopes OpenSolaris dies? With friends like that, you dont need enemies.
Right and he has ignored OpenSolaris. The company didn’t even respond to a letter from one of the leading OpenSolaris developers.
When did this become a fanboy issue? I suggested that anyone using OpenSolaris move to FreeBSD because Oracle is treating their purchase like an unwanted stepchild.
As for innovation all open source operating systems take tech from each other. DragonFlyBSD is far more innovative than OpenSolaris and more importantly ZFS and Dtrace come from Solaris. OpenSolaris is just a failed project by Schwartz who tried to lure the Linux crowd over to Sun. It didn’t work, especially when it came to attracting outside developers.
FreeBSD is a much healthier project that comes with two chief benefits of OpenSolaris.
So? Does that mean anything at all? Apple keeps quiet until they release anything, nothing is disclosed in advance. Oracle does the same thing. Apples quietness does not mean that OS X is killed soon. Sun had another strategy of being open, Oracle has a strategy similar to Apple of being silent.
Again, just because Oracle is quiet, it doesnt mean anything.
Yes, but Solaris and Sun has contributed lots, more than most OSes. Sun has contributed most to open source, than any other company. There are lots of Solaris tech that everyone uses today. I saw a huge list, but dont remember where it was. It was things like NFS and what not. The list was huge.
I dont see how DragonFlyBSD is more innovative than OpenSolaris? Everyone talks about ZFS and DTrace. What cool tech has DragonFlyBSD that everyone wants to copy? Ive never heard anything about it? But maybe you have?
Jesus. ZFS and DTrace comes from OpenSolaris. OpenSolaris is the foundation for next generation Solaris 11. All new cool tech comes first to OpenSolaris, and then, when it has been debugged, it is backported to Solaris, just like ZFS and DTrace. OpenSolaris is about innovation and the foundation for next gen Solaris. Solaris is about stability.
If OpenSolaris is killed, there will be no next Solaris 11. OpenSolaris is like Fedora. Solaris is like RedHat. OpenSolaris is the new playground and beta for Solaris 11. OpenSolaris will not be killed. Oracle has publicly stated that. Your FUD is tiresome.
How in earth do you think FreeBSD can further develop ZFS and DTrace without Sun’s engineers? Impossible. You know, there are lots of other Solaris tech that FreeBSD should port. If OpenSolaris dies, then FreeBSD will loose new cool tech, and only old stuff will be there. All new cool tech are developed in OpenSolaris or Linux (BTRFS, etc). I dont think nothing exiting is done in the FreeBSD world? Tell us of something that new and cool tech FreeBSD has done, that Solaris or Linux wants or need?
If you try to make people think that FreeBSD users are ungrateful, then you are on a good way of succeeding. Maybe Sun should stop give aid and free help to FreeBSD, maybe FreeBSD can develop a new filesystem themselves without porting it from OpenSolaris or Linux. First FreeBSD take lots of free stuff, research and development that costed years for the brightest engineers Sun has, and when done, you wish OpenSolaris to die. Nice people FreeBSD users are. Really nice. “With friends like that who needs enemies?”
They aren’t being quiet about the entire acquisition. They’re giving Solaris a lot more attention.
That was just Schwartz giving away Sun tech without any business plan. It isn’t an integral part of their company history.
Well not all the tech in DragonFlyBSD can be clean copied since it is a complete redesign of Unix. Go read about it yourself.
No they were developed for Solaris first.
Another plan of Schwartz who is no longer in charge.
You’re naive. OpenSolaris is a community project in name only. Oracle could easily can the project and move their employees that currently do all the work over to Solaris.
They’ll likely let it die through stagnation. It isn’t FUD, it’s a prediction. I also correctly predicted that Java on the desktop wouldn’t be fixed in time to compete with .net and that hiring Schwartz would be a disaster. We’ll see if I can go 3/3.
Your funniest statement so far. You underestimate the talent of the FreeBSD team but more importantly ZFS and DTrace are not as big of a deal as they used to be. ZFS is nice but has competition now and though DTrace is a useful tool it certainly isn’t required.
You seem to think I am a FreeBSD fanboy. Though I prefer it to Linux for servers my message here is that anyone using OpenSolaris should move away from it. Move to FreeBSD, Solaris, Linux, Windows Server, just about anything.
I think it is good that companies opens up. I hope you do not consider it as a bad thing as you critizice Sun’s strategy.
Thanks but no thanks. It seems that “all new innovative OS DragonFlyBSD” as you claim has no new cool tech that I have heard of. Then it probably is not worth spending time on.
Let me say this again. All new tech is put into the beta version of next gen Solaris. And only later, it is back ported to Solaris. It is best avoided to develop new critical tech directly in a stable OS such as Solaris, yes? Iron out the bugs in the beta version of Solaris, and then back port it to Solaris. This makes sense, yes? Just like Fedora and RedHat. ZFS is released first in OpenSolaris, and then backported to Solaris:
http://en.wikipedia.org/wiki/Zfs#History
But OpenSolaris has all new cool tech: ZFS, DTrace, Crossbow, etc. If the OpenSolaris code is scrapped, then Oracle has to rebuild next gen Solaris from scratch. That does not make sense, does it? Why not use all new cool tech in OpenSolaris, and then debug and release it as Solaris 11? You suggest that Oracle kills OpenSolaris, and then start development today? It will take another five years before we see Solaris 11, then!
Sure it is FUD. How in earth will Oracle deliver Solaris 11, if OpenSolaris code is scrapped? OpenSolaris is the beta version if Solaris 11. Oracle has publicly stated that OpenSolaris and Solaris will live. You are getting tiresome.
Your funniest statement so far. You underestimate the talent of the FreeBSD team but more importantly ZFS and DTrace are not as big of a deal as they used to be. [/q]
I dont see any new cool innovative tech that FreeBSD has delivered, that everyone wants? FreeBSD does not think out-of-the-box. Only standard solutions. ZFS and DTrace is not a standard solution. Unique, new solutions. I fail to see FreeBSD devs be able to develop ZFS and DTrace as Sun’s engineers.
What? ZFS has competition?? Where? Who runs a new filesystem in production at companies, and not for experimental use??
Yes, we understood your message. About everything you claim is pure wrong or FUD.
And please, if you migrate away from OpenSolaris, do yourself a favour and do not choose Windows Server. It is provenly unstable by London Stock Exchange, and will only hurt you.
You’ve got to be kidding. Oracle Enterprise Linux has a puny user base compared to Solaris, and there have been rumors of its death for the past year or so. I’d bet that it’ll be the next thing to go, not Solaris.
That must be why they proudly display their customer base:
http://www.oracle.com/us/technologies/linux/025990.htm
Along with their success stories:
http://www.oracle.com/us/technologies/linux/025987.htm“ rel=”nofollow”>http://www.oracle.com/customers/technologies/linux.html?origref=htt…
So no they are not going to ditch an established customer base.
Unbreakable Linux is used to directly compete with Red Hat. There’s a lot of server software out there that is only certified to run with RHEL. Keeping Linux makes a lot more sense than keeping OpenSolaris.
I guess we’ll just have to agree to disagree
.
That is a lot of success stories Oracle shows with Linux. But, you know what? Sun has much more success stories with Solaris.
IBM better buyer than Oracle? Are you both mad? IBM has always hated Sun.
As an analyst said, IBM would have bought Sun to get all Enterprise Unix customers, and then killed off Sun and slowly raised the prices for each new generation of technology “this new tech is much faster so a higher price is fair”. IBM would almost be the only Enterprise Unix vendor. IBM would almost have monopoly. To get monopoly is worth a lot.
Just look at valuation of big Internet companies with 100s of millions of users. FaceBook has nothing valuable except the users. If each user is worth 50 USD, then FB is worth billions of USD. But frankly, FB is nothing without users. To get access to all those users are worth a lot. If you are the only player, then the users must buy from you. You can freely set the price.
Today, IBM have almost monopoly on Mainframes and can charge whatever prices IBM want. IBM Mainframe division is extremely profitable, several billions of USD. The mainframe margin is the highest of all IBM divisions.
Sun tech was always competing with IBM tech. There was always overlap.
Enterprise Unix: Solaris vs AIX.
RISC CPU: SPARC vs POWER
IDE: Netbeans vs Eclipse
etc.
First IBM would have supported Solaris/SPARC/etc “we are dedicated and believe in both Solaris and AIX”, but in a couple of years, IBM would say that Solaris will be killed off in 5 years, and would provide an migration path to AIX. Same with SPARC to POWER. etc.
For every Sun tech, there is an IBM counterpart. IBM would be mad to develop and support similar products, that would frankly, be pure stupidity. Why would MS develop and support two totally different desktop OSes, targeting exactly the same customers? Madness. You are mad if you believe Solaris would live, coexisting with AIX. They both target exactly the same Enterprise Unix customers.
Oracle on the other hand, have almost no competing tech. Sun and Oracle complements each other excellent. Oracle will try to sell everything Sun have. IBM only wanted to kill it, to get all customers. IBM has today an outspoken strategy to shift focus from AIX to Linux. AIX development has slowed. IBM hates that Oracle is now competing. But, I tell you, competition is good for us customers. Monopoly is not. To promote monopoly is madness.
x86 CPUs are getting faster and faster. Right now, an modern x86 CPU is 5-10x faster than a IBM Mainframe CPU. Nehalem-EX reaches 70% of the speed of an IBM POWER7. Next year, Nehalem-EX will be 32nm and even faster. Then Sandybridge version will arrive, which will be much faster. POWER is not developing fast enough. It will soon be slower, but costs many many times more. AIX must be ported to x86, or it will die. Who will buy a slow CPU for a much higher price? OTOH, Solaris runs on SPARC and on x86. x86 will rule the future. Cheap and extremely fast. All OSes must be ported to x86, otherwise they will die.
Agreed. Had IBM bought Sun it probably would have meant the end of NetBeans and maybe even the death of Swing & JavaFX. OpenSolaris would have been left to rot–they’d have just taken all of the nice bits and integrated them into AIX. And Spark would have probably been sold to Fujitsu where it would die a slow, painful death.
All speculation of course–for all I know IBM would have merged their technologies with Sun’s and starting opening all their stuff up. But I kind of doubt it.
The choice wasn’t easy, but I think Sun is better off in Oracle’s hands after all.
Edited 2010-06-23 07:06 UTC
Ooo man, what are you smoking…? I don’t know if Solaris would be better in IBM or in Oracle hands and I will not comment on that. I will comment however on the technical parts of your post.
IBM has monopoly in the Mainframes? You living under a rock or something? Mainframe IS IBM product!!! Your point is the same as the point that MS has monopoly on the Windows!!!
Don’t even think of comparing Solaris vs AIX. My job used to be Unix admin @ telco company. And if SUN was competing AIX, they failed miserably. AIX is by far the most advanced Unix we have today, with a tons of advanced technologies build in, and more that you have to pay in order to use. And don’t even mention me the shit that Linux is in the enterprise ( no offense to the linux fans ).
You compare x86 CPUs with Mainframe technology?!?! Are you out of your right mind? There are those things called scalability, reliability and so on. These things are pretty much the corner stone of high-end hardware, and that’s what you get when you pay a lot of money. And that IS important for Enterprise.
IBM might have intentions to move to Linux, but I don’t see them dropping Power CPUs any time soon, or AIX for that matter.
Power is not developing fast enough?? That ain’t your gaming CPU and platform, if you want to play WoW on a good box x86 CPU is the choice for you, those evolve “FAST”, you get more cores every few months, you will have a lot to show off to your friends in school.
“All OSes ported to x86, or they die”, o man…not gonna even comment on that one.
Are you some kind of zealot or something, have you actually used AIX or even touched a pSeries box? Or are you some kind of marketing and power point wizard that loves to talk crap and throw fancy power point slides, just to show how good he is in power point presentations?
Edited 2010-06-23 13:48 UTC
Well, you actually come off as an ignorant fool.
A) Mainframe is definitely not IBM. Yes every single other company divested their resources from the mainframe, while IBM invested into the mainframes. As a result, IBM’s mainframes are the best(they are the “most” in all categories), but not the only ones.
B) Solaris 10 has some great tech, DTrace for one and ZFS as another point. Both stack-up quite good against each other. Granted, if you’re on System p, then AIX is the best option, while Linux is a very close second.
C) Mainframes application domain is very narrow. And Linux on z is the best OS for mortals to use. z/OS is plain crazy…
A) On the name “mainframe” I might be wrong, I seem to remember some article on the internet that IBM were the first to build that class of boxes and the name came from there and after sometime it became synonymous to that class of hardware. And now it’s used as general calling for, NonStop, z/Series and etc.
B) DTrace and ZFS are somewhat new technologies, the things that you see in pSeries and now slowly comming to x86 space, were basically used in the zSeries for ages and are polished over the years to near perfection.
C) True. Never used any of those on zSeries, I had read some docs on zOS out of curiosity, have to admit it’s a bit unorthodox and strange.
The point I was trying to make to the x86 fan boy is that the roomer for the death of the Power CPU, Ituanium has been around for like 10 years already. And yet they are still here ( not sure for what’s up with Itanium lately ), and there is NO single CPU to rule the all. You have different needs and hence different platforms to satisfy those needs.
I don’t think that in 2 or 3 years everyone will dump whatever they were using and opt for x86, simply because there might not be business logic for that.
Edit: Also keep in mind that Banks and other customers that use Mainframe class hardware, are not a type of customers that will dump a Mainframe box that has been working for like 20 years nonstop over night, only because the latest fasion in the CPUs is the x86…
Edited 2010-06-23 17:12 UTC
“DTrace and ZFS are somewhat new technologies, the things that you see in pSeries and now slowly comming to x86 space, were basically used in the zSeries for ages and are polished over the years to near perfection.”
Jesus. IBM has no ZFS counter part. IBM is copying DTrace now, and it is called Probevue. You are completely lost.
What I have been smoking? Why do you say that? I suggest you check things up next time, before you start to post weird things that convinces no one.
As, I said, IBM has almost monopoly on Mainframes. IBM has 90% market share. There are other vendors:
http://en.wikipedia.org/wiki/Mainframe_computer#Market
The price IBM charges is very high. Is that good for customers?
Please, next time, dont comment on “technical parts” so I dont have to spend time on posts like these.
Advanced technology? Maybe you have heard about Solaris DTrace? It has won award in Wall Street Journal
http://online.wsj.com/public/article/SB115755300770755096-R2Ct41cQ4…
Sun has won 2 Wall Street Journal awards in 3 years. Let us not talk about ZFS. I bet you have not heard about ZFS either. It is an advanced file system. IBM has nothing similar to DTrace nor ZFS. Everyone wants to copy or port Solaris tech. Actually, IBM is copying DTrace now, but it is called “Probevue”.
Regarding POWER cpu. Before, IBM mocked SPARC Niagara, saying that few cores, but very high clocked was the way to go. Niagara with many, slower cores is a dumb idea IBM said. Now IBM has released POWER7 which has many, slower clocked cores – which is suddenly a very good idea. IBM changes their mind very quick.
Regarding POWER6, it was slow despite using high clocked 5GHz. Niagara 1.4GHz won many benchmarks. IBM said “one POWER6 core is faster, therefore POWER6 is fastest CPU”. That is pure stupid. Who cares how fast a tiny part of the CPU is? The speed of the whole CPU counts. If the register, ALU, or whatever is faster, it doesnt make the CPU faster.
For instance, Sun has the TPC-C world record right now. But IBM claims they still have the record even today! Because “IBM scores higher TPC-C per core, therefore IBM has the world record”. Sure, IBM has higher score per core, but please look at TPC-C and see who has the world record. It is Niagara 1.4GHz machines.
BTW, IBMs TPC-C world record machine, a Unix P595 costed 35 million USD list price. With discount, it costs 17 million USD. That is a ridiculous sum. How can you charge 17 million for ONE AIX server? Sick.
According to wikipedia on “FUD”, IBM was the first company to use FUD. Never trust anything IBM says or you will be brainwashed. For instance, IBM claims that POWER6 has bandwidth on 200GB/sec – but IBM has added all bandwidth in the chip! That is wrong. If there is a bottleneck of 1GB/sec, then it will never be faster. You can not add bandwidth. That is pure lies.
IBM advanced technology? Yeah right. IBM just follows Sun and copies everything Sun does. And also IBM lies or distort the truth.
Agreed. Linux is shit in the Enterprise. To unstable and buggy and doesnt scale well.
How about checking facts before posting? I am talking about Mainframe CPUs. “They are dog slow”. You can emulate a mediumsized Mainframe on a PC, in software. Sofware emulation is a factor 5-10x slower. If x86 could run native speed, then you would need ca 10 modern x86 to match the biggest Mainframe with 64 cpus:
http://en.wikipedia.org/wiki/Hercules_emulator#Performance
The biggest Mainframe with 64 cpus, gives 28.000MIPS. Eight modern CPUs gives 3200 MIPS under software emulation. If run native code (5-10x faster), those eight x86 would give 16.000-32.000MIPS.
The TPC-C Unix machine with the world record, costed 17 million. But mainframes are much more expensive. You figure out the price on the biggest mainframe. You could buy a couple of PCs to outrun it.
AIX development has slowed. IBM has officially outspoken intentions to migrate AIX to Linux. IBM has said that. Do you really believe POWER cpus will live when AIX is dead? Intel CPUs are closing in, develops very fast. And much much cheaper. Why would IBM spend money on expensive POWER when Intel is cheaper, and soon faster?
Intel and AMD spends huge money on R&D. Soon Sandybridge will be here, next generation which is “the biggest step ever”. Where is POWER8? It will take several years before we see POWER8. And then intel will have released two or three new generations of x86. POWER will be slow, but schweine expensive. Why would you buy one expensive, slow POWER machine? Look at Itanium, it slow and everyone abandon it. Which is a pity, because competition is good for everyone.
Of course you are not going to comment that. But.. wait, you did comment that. You said “no comments” – which is a comment.
Anyway, IBM is going to Linux. IBM will abandon AIX. IBM has said that. Officially. So, I am right: AIX will die. And it is not ported to x86. And POWER will also die, when AIX dies.
I wonder, are you the IBM zealot? IBM has promised not to sue companies that use 511 patents that IBM has released. Now, someone is selling the IBM Mainframe emulator “TurboHercules” and guess what? IBM is sueing him. Hilarious. According to wikipedia, IBM was the first company to use FUD.
Solaris has shipped more than 20 million licenses today, including OpenSolaris. There are three Solaris machines shipped for every AIX machine today. Look at the forums, the OpenSolaris part is quite big. AIX is miniscule or dont exist. AIX is a small niche, shit load expensive OS. No one can run it at home on x86. OTOH, Solaris is very common in Finance and Telecommunications.
And, Oracle Enterprise Database runs on Solaris, more than any other OS. Oracle has 75% of the database market (including MySQL). IBM has 15-20% of the market, with DB2. Oracle database is developed on Solaris, and then ported to other OSes. Solaris is the premier choice for ORacle database. AIX? No thanks. Who wants a soon-to-be-dead closed OS?
You forgot to mention that IBM used 1 machine, and Oracle used a cluster of 12. Even TPC distinguishes between clustered, and non-clustered. Clustered is just plain silly, and there are only 2 in whole history of TPC-C.
That’s wrong. The prices in the disclosures include every piece of hardware and software used in the benchmark. That includes cables, switches, x86 machines to generate the traffic, etc.
TPC-C makes it really easy to look at the prices. There’s a nice table in the executive summary. You don’t need to look through those huge full disclosure reports that are hundreds of pages long.
I saw a few import things when comparing the 2 submissions.
http://www.tpc.org/results/individual_results/IBM/IBM_595_20080610_…
http://www.tpc.org/results/individual_results/Sun/Sun_T5440_TPC-C_C…
If you look at the table for the p595 submission, most of the money was spent on storage. The list price, before discounts, was $21 million for storage. The list price for the p595 was only $12.6 million. I think a 50% discount is reasonable since they gave a 55% discount on the whole package. So a p595 is more like $6.3 million.
Why did IBM spend so much on storage? It was June 2008 when they ran this test. SSDs weren’t an option. 1.5 years later Oracle was able to use some SSDs and reduce their storage costs. TPC-C has always been heavy on storage. Even with the SSD advantage, the Oracle list price for storage was almost $13 million.
The other big cost is the DB software. Oracle RAC is quoted at $7.8 million for a 3 year lease. That’s right, like a car, when time is up, you have to pay more if you want to continue using it. And no, I’m not talking about maintenance here. At least the DB2 is $2.3 million and you get to keep using it.
Ok, I was wrong. The IBM P595 server which hold the former TPC-C record, did not cost 17 million USD because the cost included:
A) the server
B) extra hardware such as RAM and discs and cables
C) 128 client computers to generate load
D) third party software
So, if we omit the client computers and omit the third party software, and instead only look at the top configured unix server and necessary RAM and cables – then we dont see a price of 17 million USD. The correct price is instead 16.5 million USD.
So you need ONE unix server for 16.5 million USD to be able to reach that high TPC-C numbers that IBM so proudly boasts. Isnt that sick? Almost 17 million USD for ONE top configured unix server? Imagine the price of one top configured IBM Mainframe which are much more expensive than a lousy Unix machine. And the Mainframe is beaten by one x86 server with eight modern Intel/AMD cpus in terms of processing power. And people thinks monopoly is good for the customers? Pay extremely much for heavily overpriced, slow IBM gear? Are people mad?
Regarding the current TPC-C record. Yes, Oracle has the top spot with a cluster of twelve Niagara SPARC machines. The question is: how can IBM still claim they have the current record because IBM used fewer cores? I dont really get it. Can you explain this to me? Or, is it pure IBM FUD as usual?
IBM was the first company to use FUD and are FUD masters:
http://en.wikipedia.org/wiki/Fear,_uncertainty_and_doubt#Definition
I advice everyone to be very cautios when IBM claims something. Before Microsoft, IBM was always the big Evil company. MS has changed strategy and is nicer now. Not as much FUD as before. But IBM has never stopped lying. They are still FUDing a lot.
For instance, IBM claims that one Mainframe can virtualize 1500 of intel x86 servers. That sounds magnificent. Until you look up the facts. IBM assumes that all x86 servers idle and the Mainframe is loaded 100%! If this is ok, then I can claim that my old laptop can virtualize twenty Mainframes (they must idle). I can software emulate Mainframes on my laptop, that is true. But everyone thinks I am lying if I claim that my laptop can virtualize twenty Mainframes, right? What happens if the Mainframes must do some work? Then my old 1GHz Pentium M laptop would crumble. So it is clearly a lie from IBM. We have all seen my links that proves that a x86 cpu is much much faster than a IBM Mainframe CPU. An IBM Mainframe can not virtualize few x86 servers that do some work, let alone 1500 of them! Lies and FUD.
Likewise, IBM states that the new POWER7 servers can virtualize 800(?) intel x86 servers. But if you look up the facts, it is the same things. The x86 servers in the comparison are very old, and they all idle. And the POWER7 server is fully loaded. This is clearly another IBM lie and FUD. What happens if the x86 servers must do some work?
Master of FUD and lies = IBM. Dont trust IBM.
The 8 core / 16 thread Nehalem Xeon is just a dream when it comes to running games and doing homework. The 24mb of cache was built specifically for Crysis. Seems a tad noisy though.
Seriously though x64 is going to dominate all server sales, including the ultra-high end.
Why does a network company buy a server company anyway?
When I worked at a big name company they used to buy smaller companies like Sun just to provide in house equipment. It was cheaper to buy the whole company and outfit the big business and then basically throw away the smaller used up inventory company.
So does Oracle really need a bunch of in house servers or some programs that Sun has? I doubt there is any business reason based on sales that makes this deal work otherwise.
It is not cheaper to buy a whole company just to get some hardware for internal use. It is much cheaper to buy commodity gear off the rack. Then you dont have to support nor develop and research.
Regarding Oracle need of Sun. Oracle’s big Enterpirse database is developed primarily on Solaris, and then ported to other OSes. Since many years, Oracle database is also running on Solaris, far more than any other OSes. In fact, Solaris is the prefered choice to run Oracle database. This was outspoken, long before Oracle bought Sun.
Oracle and Sun is a very good match. IBM has got a new tough competitor. Sun tech is arguably more superior than IBM tech. Also, Oracle database are arguably more superior than IBM databases. Oracle has now ca 75% database market share (including MySQL). IBM has 15-20% market share with DB2.
Therefore, I see good value in Oracle buying Sun. And also Sun and Oracle sees it. When IBM tried to buy Sun, no Sun owner wanted that. Then Oracle came as a saviour and IBMs offer immediately got rejected.
Oracle + Sun will give IBM a good run for the money. Us customers will benefit from this. It is good for us, for Oracle, for Sun and bad for IBM (because that forces IBM to try to stay competitive and develop new tech. Otherwise IBM would become lazy and charge high prices for nothing)
Again, IBM would have killed off everything that Sun has, and left would only be AIX. That Unix is not opened and you can not use it at home. It only runs on schweine expensive POWER servers. IBM is very closed. Solaris would have died, and AIX is out of reach. Left to us customers is Linux and FreeBSD. Is that good for us customers? No. Pure madness to believe that IBM would have been good for Sun.
I’ve hated the news of Sun being acquired by Oracle the moment they wouldn’t let me migrate my Sun Developer account unless I gave my company’s name, address, phone number, etc.
Seems plenty of activity when the devs are around on #opensolaris and the CIA notifications. PSARCS are still being opened and closed, Nexenta is still busy on it, ZFS is moving fast.
For a ‘dead’ and ‘failed’ project theres an awful lot of Oracle resource still working on it
One thing I have noticed is that the lx-branded zone just came out of OpenSolaris (ran linux 2.4 kernel bins) – perhaps because Oracle are tweaking that inhouse to migrate Enterprise Linux users to Solaris Zones and don’t want to give that away for free?