Slashdot Mirror
Oracle Joins IBM AIX Collaboration Center
pgsqlDao writes "CRN is reporting that Oracle is joining IBM's AIX Collaboration Center. 'IBM announced the center Dec. 16 as a $200 million investment where it will centralize AIX development, customer relations and advanced features for independent software vendors. While the figure represents existing salaries and equipment drawn together under one roof, it also represents some shift in emphasis by IBM from Linux back to its mature Unix operating system.' In November Oracle announced that it has chosen Solaris 10 as it's preferred development and deployment platform for X64 computing."
it also represents some shift in emphasis by IBM from Linux back to its mature Unix operating system
This isn't necessarily a shift. Linux is perfect for many, many applications but there are a lot of applications and installations out there that are still relying on AIX. Even Linux users should be heartened to hear that IBM won't abandon their customers lightly.
Look how long they supported OS/2!
A monster ate my homework!
SCO has been trying to argue that by promoting Linux, IBM has killed off its Unix business. They want billions of dollars in damages. This development shows that both IBM and Oracle don't think Unix is dead. SCO's business is dead because they have the bad habit of suing their customers not because Unix is going extinct.
Just a year ago, the preferred direction as declared by Oracle was Linux boxes as a "grid" building block. IBM was pushing Linux. Sun has been schizo regarding x86 and ultrasparc and Linux support for a few years now.
AIX service and applications are more profitible for IBM than are Linux services. I would bet that shoring up AIX is in repsonse to Sun's greater emphasis on Solaris. Both are formidible for large enterprise applications, but Solaris is now open source and picking up steam. This will be interesting to watch as two giants duke it out.
Who says non-Linux UNIX OS's are dead? Far from it.
the reason why Oracle went with Sun is because Sun was basically supporting OSS as well. Oracle is just as panicked about OSS databases as Sun is about OSS OSs. And while IBM can lose AIX, Sun can lose Solaris, Oracle will die if oracle the DB dies.
I prefer the "u" in honour as it seems to be missing these days.
Aside from GNU utils (which AIX does have packages for), what on earth does Linux have that's 'way better, in all respects' ?
As a long-time sysadmin for AIX and Linux, I find this laughable at best. AIX has so many enterprise-level tools and resources that are so well-developed, that Linux, IMNSHO is still far FAR behind.
So please, enlighten us.
Is that a salami in my pants or am I just happy to be me?
> Having worked with both, I must say that Linux is way better, in all respects.
What are you comparing? AIX 4.3 to RHEL 3?
AIX is now on v5.2 & v5.3.
I find that I keep my db2 servers on aix even though db2 also supports linux. The better logical volume management, firmware reporting, etc makes it a slightly better platform. I haven't benchmarked the two but would expect AIX to be the fastest. And perhaps most importantly - the fewer patches means that it costs far less to support.
Why did you find linux better? Even the gnu tools are now commonly installed on aix.
ken
IBM will gaurantee a transaction level with Oracle on AIX that they say they may gaurantee at some point in the future on Linux. I don't think this is just marketing to keep selling AIX, I really think it is still a better platform for a certain level of service.
It's hard to believe that's how Micronians are made. Why don't we see it right now by having you both kiss one another?
... Keep your friends close, but your enemies closer.
And "it's" is not a possessive pronoun.
I admit I've never used LVM, in either Linux or AIX. Considering the price of hard disk storage these days, it's much easier to buy a new, bigger disk with the needed physical capacity than to manage an old array of disks.
And if you mention the cost of buying a new disk, I must answer with the much bigger cost of buying new software. What people often fail to consider when comparing commercial systems with free/gratis software is the effort needed to justify acquisitions. When I need a new package in Linux I either "apt-get install" or "./configure; make; make install" it.
I don't need to go to the lengthy and troublesome process of filling forms, attending meetings for the next year software budget, editing spreadsheets, etc, etc, that's inherent to commercial software.
Had you used LVM, you'd know that it's more than just a matter of buying a bigger disk when you run out of space. I can't imagine using a system without LVM now. All of my Linux boxen (both desktop and servers) use LVM. Curiously, some of my production Solaris servers don't (for our specialised applications, it's not necessary), although we do where appropriate -- DB servers being the obvious example.
"The invisible and the non-existent look very much alike." -- Delos B. McKown
Be careful bragging about AIX, it's not at the head of the UNIX class. Solaris 10 kicks AIX butt in features and performance. Solaris10 is also Open Source and runs on X86 boxes all the way to E20K 72CPU machines. AIX is not free, Solaris 10 and Linux are free to users (maintenance is extra, unless you chose to get it via the LUGs or other sources) Linux is a very good choice for a Small/Medium Business environment and even for some desktops. Right now Linux does not scale really well once you get past about 16 CPUs.
Another big win would be to replace the generally crappy packaging systems with something like apt. A few companies have made a stab at implementing package systems that work around dependency hell, but I've never encountered one that works as well as apt does. And I'd sooner dig my eyeballs out of my head with toothpicks rather than work with SMIT ever again...
Those two steps alone would make commercial unices a lot nicer to deal with. While the other UNIX variants may be more mature than Linux in the kernel department, they are DECADES behind in the user interface arena. OSX being a noteworthy exception.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
OK. I'll byte: what is exactly the advantage in LVM? What is it that LVM, in either Linux or AIX, provides? If it's a simple matter of growing capacity, why isn't a bigger disk (very cheap) better than spending any time at all in configuring LVM? For more complex applications, wouldn't RAID be better?
Better compare that to RHEL 4 instead of 3 while you're at it :-P
> Better compare that to RHEL 4 instead of 3 while you're at it :-P
dang! i hate mistakes like that.
> Solaris 10 kicks AIX butt in features and performance.
Do you know of any public benchmarks that would show this? I know that AIX 5.3 on Power5 is almost 50% faster than AIX 5.2...
And more to the point - IBM is much healthier than Sun, and much more likely to be around in 10 years.
> AIX is not free, Solaris 10 and Linux are free to users (maintenance is extra, unless you chose to get it
> via the LUGs or other sources)
That's not really accurate. You *can* get a free linux distribution, but quite often you'll find that applications require RedHat or Suse. And RHEL4 is more expensive than AIX.
> Linux is a very good choice for a Small/Medium Business environment and even for some desktops.
Sure, I'd actually recommend both. Want to run websevers or print servers? Linux works fine. Want to run critical app or database servers? AIX is a better fit. Obviously linux will be better on desktops or laptops.
> Right now Linux does not scale really well once you get past about 16 CPUs.
That seems to be a diminishing issue: the big driver of massive SMPs was Oracle, and they've headed towards a more distributed architecture. It certainly doesn't matter at all for db2 or informix - where they have for years supported more of a beowulf-like clustering approach.
Your joke aside, the 200 million being spent isn't new money being spent on AIX. It's really just existing resources being pooled together under a new name. Just a marketting ploy. From the article:
Open Source Java DAO Generator
"While the figure represents existing salaries and equipment drawn together under one roof, it also represents some shift in emphasis by IBM from Linux back to its mature Unix operating system."
How do you figure that? If all it is is shifting the books around, I don't see how it's changing emphasis on anything. Sounds more like simply reorganizing in the name of effenciency.
- It's not the Macs I hate. It's Digg users. -
I don't think any of these companies IBM, HP, Sun has made the same level of progress linux has in the past 3 or 4 years. Maybe the massive layoffs had something to do with it? Maybe the extreme outsourcing?
Yes, there are quite a few on Sun's web page. Look down at the bottom of this page: http://www.sun.com/software/solaris/index.jsp Sun has over $5B in the bank, they are not going anywhere. They are the largest supplier of computers to the US Gov't. Comparing Standard plans, RHEL is about $799/yr maintenance. AIX is $299. (IBM web site). Solaris 10 is $240 (Sun web site) However, there ARE a lot of small shops that will use the 'Net as the source for Maintenance. RIsky in my opinon. Linux works very well as web/print servers. Lightweight tasks it handles well. Actually SMPs still are hot, there are problems where it fits well. Oracle runs just fine on SMPs or on distributed. You really need to know more about HOW a company uses Oracle to tell which way they should go. Beowolf clusters where all the machines are connected in a "network" and the scheduling is managed centrally. That's good for problems that can be broken into parallel threads that can come back together (synchronize) as needed. I can't think of a General Purpose business problem that can be that parallelized to take advantage of this approach. There are certainly problems in Science that can be done this way. Genetics comes to mind. Graphics (high end rendering) can also benefit.
Check out Solaris 10 and other new products from Sun before you assume that. Also, Linux is a WORLDWIDE development effort, which obviously has folks working on something 24x7. Other software products don't have this "staff" for development. There is still a lot of corporations that do NOT embrace (and will not embrace) the open-source concepts of Linux. So that limits Linux use in Corporate computing. However, on the desktop, Linux is GREAT. It kicks Windows butt.
Linux had a lot of catching up to do.
Serving Suggestion: Defrost
At their current rate, probably 20 years :-)
They almost broke even last year.
Take a look at the unix03 register. Only Solaris 10 and AIX are Unix03 compliant. That's why it makes perfect sense for IBM to continue its commitement to AIX (and for Sun to Solaris of course).
Linux (I mean Linux in the broader sense, not just the kernel) should strive to achieve some sort of formal compatibility to a standard like that.
What the hell is AIX? I've been in IT for 10 years now and never heard of it.
HP-UX has way more enterprise level tools than AIX could ever offer. And most of these tools are either already ported to Linux or will be ported in near future. Ever heard of HP Serviceguard? No buddy, it's not something you can eat.
Your comment "I find this laughable at best" makes _you_ look laughable to say the least.
"it also represents some shift in emphasis by IBM from Linux back to its mature Unix operating system."
IBM will support OSS including Linux as long as the PR value exceeds any investment and not a second longer. Since they lost leadership of PC development, IBM has never met idea or consortium that they didn't like, but in the end their participation has little impact on the adoption of the technology involved.
Moron...we're talking about Linux vs. AIX...and no, I haven't heard of Service guard, but unless it comprises:
...then I think your point, however offtopic, is still inaccurate.
- robust LVM+mgmt tools (which sistina's LVM is not yet)
- HACMP-level clustering+mgmt tools
- true virtualization of hardware including dynamicity and scheduling
- and oh, the other zillion other things that AIX has that makes managing it a dream.
Is that a salami in my pants or am I just happy to be me?
kpharmer said: Sure, I'd actually recommend both. Want to run websevers or print servers? Linux works fine. Want to run critical app or database servers? AIX is a better fit. Obviously linux will be better on desktops or laptops.
The "print server" comment sounds like some FUD from the late 90s.
Do you have any evidence to support your statement that AIX is somehow better than linux for databases or critical apps? Sure, you can use AIX for those things - but linux is certainly a viable choice. With 30 million customers and 24x7 operation, amazon.com seems to indicate that linux does a fine job powering web services and back end databases.
AIX somehow "better"? Sure it's mature and stable, but it is showing its age, and it is a bit eccentric. And note that IBM is careful never to compare AIX performance against a 2.6 kernel on similar hardware - but they are always quite keen to show the scalability of AIX as compared to the obsolete 2.4 kernel.
BTW I see we are still trotting out the old "low end x86" spiel - that is really tired, BTW. SGI will sell you a 256 CPU, 64 bit linux server right now if you've got a million bucks.
And what about the fact that linux is dominating the supercomputing space? can anyone really imagine that we are still talking about "low end x86"?
I think some dinosaurs see their way of life threatened by linux, so the dismissal of linux' capabilities is a perfecly understandable psychological defense mechanism. I trust that time will teach these folks a lesson or two.
IBM is a huge company, the IT industry is less stable than for a long time, and they are putting their eggs in more than one basket because they can?
Virtually serving coffee
> Do you have any evidence to support your statement that AIX is somehow better than linux for databases or critical apps?
Sure, it's the reference platform for db2, has a great logical volume manager, and has far fewer patches than linux. Note that if I was running a lot of critical oracle I'd be leaning towards Solaris for similar reasons. Note also that the difference in patch frequency can result in a cost difference of a few thousand dollars a year in admin labor.
> Sure, you can use AIX for those things - but linux is certainly a viable choice. With 30 million customers and 24x7
> operation, amazon.com seems to indicate that linux does a fine job powering web services and back end databases.
You're confusing issues. Of course you can support mission critical applications on linux. But that doesn't mean it is the easiest and currently best way to go for everyone.
> AIX somehow "better"? Sure it's mature and stable, but it is showing its age, and it is a bit eccentric.
It's always been a bit eccentric. As linux gets more popular AIX will have to change to keep up - you can already feel minor differences with simple utilities falling behind - like cron (doesn't support time equations, etc). But since gnu utilities are now provided on aix it is helping. And I can generally live with these issues in return for the mature and stable parts. Especially for a database server.
> And what about the fact that linux is dominating the supercomputing space? can anyone really imagine that we are still talking about "low end x86"?
Linux is now dominating the supercomputing space when it comes to MPP architectures - machines that can be built from a highspeed internal network along with thousands of share-nothing nodes. First off, this architecture has existed almost unchanged since 1994 or so. Secondly, this only really helps satisfy some supercomputing needs - where you can divide the problem nicely into separate pieces. Third, it's a simple matter of economics. Note that this architecture was initially developed for super-computing, but then sold most of its units for large-scale databases. Where it can make more sense to use a two-way P5 than a two-way opteron.
But today when I need to support a financial database on a pair of redundant servers, I'm going to heavily lean towards aix or solaris (assuming db2 or oracle). What works in an MPP environment doesn't apply here. This will probably change in five years, but for now it holds.
> I think some dinosaurs see their way of life threatened by linux, so the dismissal of linux' capabilities is a
> perfecly understandable psychological defense mechanism. I trust that time will teach these folks a lesson or two.
I think that it's unfortunate that some linux advocates are unable to accept the idea that any OS can be better at anything than Linux. Even admitting that linux will probably surpass unix in five years isn't enough - they demand that you admit that it is in all ways better than unix right now.
There's really no room for dogma in IT.
SCO's unix biz was terminally ill long before they started suing people (although that didn't help) The law suits were simply an act of desparation. SCO is dead because their product lacked all the high-end features found in most of the big 64-bit unix flavors like Solaris, HP-UX, and AIX. SCO was living in the the low-end 32-bit intel area of the Unix neighborhood and they just couldn't compete with Linux there and 64-bit Solaris/AIX RISC on the high end. BTW... I heard that SCO's biggest customer (McDonalds) was switching to WinXP for their point-of-sale equipment. Probably the last nail in the coffin.
kpharmer says: I think that it's unfortunate that some linux advocates are unable to accept the idea that any OS can be better at anything than Linux. Even admitting that linux will probably surpass unix in five years isn't enough - they demand that you admit that it is in all ways better than unix right now.
Yes, that is unfortunate - or would be I suppose, except that I've never met any linux users with the mindset you describe. We do feel, though, that this supposed gap in (performance/stability/enterprise-readiness) between linux and it's elder brethren is a bit overstated.
It's common knowledge that Linus will not tolerate any kernel modifications that increase performance on high-end machines at the expense of the common case. This may not be the absolute best decision for the 32-way database server, but it makes good sense for the vast majority of usage scenarios. Unlike AIX, Linux will never be optimized strictly for high-end hardware - Linux doesn't have that luxury, nor that narrow a niche.
Even though it must perform well over a broad range of hardware scale and capability, that doesn't mean that Linux is neccesarily a slouch on high-end hardware. It just means that kernel modfications to enhance high-end performance must be clever enough that they don't harm low-end performance. This requires more skill and more effort, and it takes longer for high end performance enhancements to make their way into the main tree, but the results are worth it in the long run.
Ill gladly concede absolute optimization of Linux performance for extreme high end hardware, if it means that it performs respectably in 99% of real life situations.