Oracle has announced that it has stopped development for all its software on Intel's high-end Itanium server processor. Oracle works on its own cycle and does what it wants - which is why it waited until a week after Leo Apotheker's coming out party as president and CEO of Hewlett-Packard, and until 8pm Pacific time when …
...trust el reg to come up with an entertaining title to spice it up a bit.
HP's Intanium Systems have no Integrity
Notice how Intel was very careful not to commit to the Itanium chips....just the HP operating systems.
Paul Otellini, president and CEO of Intel Corporation “We remain firmly committed to delivering a competitive, multi-generational roadmap for HP-UX and other operating system customers that run the Itanium architecture.”
Matt B....we will miss you
It's the same thing as the only platform HPUX ships on is Itanium-based Integrity. If I were an Oracle customer (I'm not) this would truly piss me off. Most of them have to feep like their being herded down a cattle chute about now..
Timing is everything
Today was HP's annual stockholder meeting. Oracle's timing is Larry and Mark's ultimate payback FU to HP.
I warned a while back that this was the last thing I wanted Oracle to do - break their DB product. I don't see how Larry thiks this will benefit Oracle, if there's anything we couldn't move onto Xeon then it would be going to Power, not SPARC in any form. This does remind me of the kerfuffle around Slowaris x86, where Snoreacle said they were dropping hp's license to support Slowaris on ProLiant, then a few weeks late it was all back on when hp stumped up a larger license fee. Maybe this is just another extortion attempt by Larry, but it just makes him look rather desperate if he thinks this is the only way he can sell SPARC servers!
You can see why he didn't try it with IBM - they have DB2. Maybe it's time for hp to buy a database company. Of course, the fun starts if Leo goes and buys SAP and a database company, then drops SAP support for Oracle, then I'd really have a ton of migrating work to do! Either way, I expect I'm going to have a busy year trying out the options. As usual, the last person Larry thinks about is the customer.
".......the only platform HPUX ships on is Itanium-based Integrity....." Try a little reading, there's both OpenVMS and NonStop running on Integrity, and both are often used with Oracle.
Um.. the NonStop is *based* on Integrity systems and OpenVMS runs *on* Integrity. No additional reading on my part required..
Is it just me or is Oracle painting itself into a corner ?
It's not like this will get HP to invite the Oracle Sales rep to a joint session on their accounts where they before might have teamed up against for example IBM or SAP and Fujitsu or...
The word Megalomania comes to mind.
Why would they want to partner with HP? They have a full stack of hardware & software they can flog to customers themselves, they don't need another partner.
You do realize that Oracle now owns Sun and this includes Sun's hardware business?
Do you honestly think that if HP invited Oracle to the party that the Oracle shark er sales rep wouldn't hesitate to try and sell Sun hardware in to the deal?
Most of us do not need what Itanium offers. Some corps do though.
Well, considering the majority of the old Oracle reps don't have a clue about hardware, and the few ex-Sun ones that made it alive into Oracle only seem to have done so because they didn't have anywhere else to run to.....
Guys you do know..
That there is such a thing as history.. just cause Oracle got SUN's HW business doesn't mean that HP hardware sales reps' haven't spend quite a lot of years before that going to marked together, for example to try to replace a Mainframe installation with DB2.
Now, I have yet to meet a Oracle sales rep who had just the faintest clue about HW, so what is his options ?
What you guys aren't realizing is that Oracle also just raised the price on Itanium, so the sales rep is more than happy to sell Oracle on HP Itanium. And now he has to compete with not just SAP and IBM, no he has bloody the whole industry after him.
And sure you say that he can just push a Exadata solution. Sure.. yeah but that is a tough sale, it's expensive and cannot really be used for anything else than an Oracle solution stack.
And a friend of mine who's an Oracle consultant (doesn't work for Oracle), wasn't to fond of the support contract terms.
Just not that joined up.
We're trying to get Oracle to do a combined Hardware/Software deal, it's like try to pull hens teeth. We are still being offered a better price using IBM hardware and Oracle Software, go figure.
But then IBM global services sell more Oracle software than IBM.
The only way they'll do it is with the database, middleware and desktop appliance servers.
Oracle are not easy to business with, even £60m's worth.
...in the coffin of HP-UX. I suspect the majority of HP-UX deployments have been for Oracle database servers and essentially dead-ending the platform as viable will move customers off to some other platform, but whether that is Solaris on SPARC, AIX on Power or Linux on x86 is anyone's guess.
Oracle are busy alienating people around the world by dropping support for various products; it's been pointed out to me that ASMlib on RHEL 6 has been dropped unless you have an unbreakale linux sub - this is presumably designed to "encourage" users to switch to Oracle Linux.
Surely x86 will be seen as the safe bet?
RE: Linux ++
Well, maybe not, at least not for Larry. The question that's got everyone thinking here is how far is Larry willing to push it? Suppose he doesn't stop at Itanium, or even Power, what if he gets really uppitty and starts pushing CMT by either bumping up the license costs on x64, whether it's on Linux or Windoze? Could it get to the point where Oracle is only available on SPARC in some form, or x64 only if it's Larry's Untakeable Linux or Slowaris, and you have to like it or lump it? That's not an idea we like, we got badly burnt in the past by being too dependent on Sun so we like to keep options and choices available. Suddenly the question becomes how hard would it be to unplug Oracle, and the answer is painful but not impossible.
Time for a rename?
It was a valid term when the world was dominated (at least in crunch-ability, if not numbers) by the mainframes, with their multi-kilowatt power consumptions and their multi-multi-kilowatt cooling requirements and processing MIPS-age to boot. Nowadays our "micro"processors are at least as capable as the CPUs of 30 years ago - even as capable as the mainframes of 10 years ago.
Isn't it time we reserved the term for the _real_ microprocessors: the PICs, AVRs and other chips that still measure their clock speeds with MHz as opposed to GHz and just started to refer to the devices in our computers as "processors" again? They've earned that much
Stopped using the term Microprocessor when I started dealing with the 68000, it really doesn't apply to anything beyond 8 bit computing and the earlier 808x range (8086, 8088). If you want to be picky and what is a processor and what is a microprocessor, maybe an on-wafer FPU should act as the deciding factor...
Depends on why you think they're called *micro*processors. If it's because of the scale of the individual transistors on the chip, maybe we should be calling them nanoprocessors now, since the process size used in processor chips has shrunk over the last 30 years or so from microns to nanometres.
Microprocessors are a processor on one chip...
...so Itanium, Nehalem, the various SPARCs, etc., etc., are microprocessors.
(Interestingly, many POWER processors are NOT, as some are multi-chip modules - and not just multi-chip for multi-core, but different parts of the cores on different chips.)
So, Larry says "Sink the Itanic" ?
This doesn't just affect HP-UX - there may be some OpenVMS shops out there running Oracle on Integrity, though I guess they'll be stuck on 10g.
End of the road for RDB?
Oracle RDB (nee Digital RDB) only runs on OpenVMS and OpenVMS only runs on Itanium, so it appears that product line wil be wound down.
I can see several outcomes
If you are a software house would you use Oracle's database as part of your application? Which would in turn limit the hardware you can run on, potentially limiting your sales? Probably only if you want to grow large enough to be acquired by Oracle . . . So real innovation will use an alternative platform which is likely to slow their acquisition trail in the longer term for them to become a dynosaur.
Would you buy an IBM P-Series server now to run your Oracle ERP and increase the probability of your strategy only having a five year life . . . if you were IBM would you invite Oracle into your accounts? So Oracle might initially grow more rapidly in the short term and then be examined by the US legislature as they will have become too dominant - much as nearly happened to IBM and Microsoft and did happen to Standard Oil, Bell Telephones . . .
Problem is... what do you choose instead? Besides, lack of platform choice doesn't seem to be hurting SQL Server much. The way things are panning out, RDBMs will be reduced to
SQL Server => Tied to Windows Server, or Windows Desktop for the 'lite' version
Oracle => Tied to Solaris, or Linux for the 'lite' version
Or you can go open source and hope that the features you need are implemented.
Maybe this time intel will accept that VLIW is crap for general purpose CPUs running typical software. How many times do they have to make the same mistake? Too bad the itanium took out so many better CPUs with it (the alpha, the high end MIPS (although MIPS seems to have survived), I don't think pa-risc was ever great, so who cares about it). Just about took out SGI for good too.
Trust? or Anti Trust
Oracle having grown fat for years on the back of selling it's SW on HP's boxes might find that it now faces anti-trust cases around the world for having used those fat profits to buy a competitor and then trying to cup the old golden egg laying goose out of the party.
In the mean time, endangering VMS is likely to result in the US Navy thinking Larry's latest yacht would make a good bit of target practice.
Not until they have a de facto monopoly on database sales. But damn close, none the less. Or you could argue the other way - maybe HP should be shipping alternative versions of these Itanium based servers using other processors... eggs, baskets and all that...
Does the US Navy hire legionaires ?
If that gets one to have a few shots at Larry's yacht it'll be worth going through the drill. Can't be any worse than working for Oracle, anyway. Sign me up !
The main problem with Itanium
The main problem with Itanium is that there are only 3 vendors to choose from: Intel, Intel and Intel (oops! on a second look, it seems it's just one).
Oh, and the thing is patented to death and heavily proprietary, so forget about any competitors jumping on board later. So much about choice.
Who is going to invest in THAT, knowing that if Itanium ever gets off, Intel will ramp up prices to the sky?
The "difficult architecture, lazy programmers" card Intel has been playing since the launch of Itanium is moot. Just give those programmers a good compiler (which has already happened), and they will happily program for ANY architecture.
WTF? *If* Itanium takes off ???
You know it's 2011, right?
Did you write this 10 years ago?
Oracle isn't going to abandon Power
That is, of course, unless they want to lose all of their big iron clients to SAP and that business isn't insignificant.
Since the Itanic never caught on with anyone other than HP, I can see some reason for the move.
shitty thing to do
Not only giving HP a shoeing, but also making Intel's announcements for them. Big business really is a forest of bastards, isn't it?
This has all the hallmarks of manipulation. Whatever Intel's policy on the itanic (kudos!) was, I bet that by tomorrow it will be whatever Elison says it is.
I feel a bit like I did when Dr Beeching closed the railways. Let down, mourneful, sad and full of rage. For Glod's sake, surely diversity is good, not bad? Who is going to do a Flanders and Swann for cpu architectures? http://www.youtube.com/watch?v=U6OHD2uCpfU
Insane or Genius?
I know a lot of very large shops that run an Oracle HP-UX combo, not good news for them. The anti-trust is an interesting angle, wonder how long it will take for them to do it to Power as well....
No if only someone could bring to market a database tech that completely spanked Oracle that would liven things up.
I also think you'll see a lot of the smaller IT shops moving towards SQL and Opensource as well rather than pay the Oracle tax which is becoming worse that UK Fuel duty
Support is very costly
It's pretty simple really. Each different platform on which you support your software costs money: you have to compile and test it, and create installers, and documentation sets, and then you need to create and distribute patches and retain teams of support staff in each global region for the extended life of the product, so for products launching next year this will take you towards 2020. If you only have a few hundred sales, or even a few thousand, it will cost you way more than you will make in sales.
With Exadata, P-series and SPARC at the top end, and Wintel and GNU/Linux at the bottom end, the potential market share for Itanium based systems is pretty small. For Intel and HP it is a blow, but for Oracle it is sensible housekeeping.
I suggest you take a look at the latest matket share figures
SPARC revenue is now significantly lower than Itanium revenue - based on that argument if Oracle was really going to end of life one of its development platforms, it should have been SPARC.
No, this is Larry saying, I can't win fair, so I'll win dirty... ANYONE who trusts Oracle to do the right thing by them now is on to a loser, when Larry said he wanted to be like the old IBM, he meant becoming a monopolist...
Time to look at Sybase, Ingres, DB2, Informix - basically anything but Oracle
RE: Support is very costly
Dunnie, your argument doesn't hold water. For a start, the majority of the task is the initial code base for UNIX, which should be pretty common for AIX, SPARC-Slowaris and hp-ux. The real work is done later when the code is optimised for each platform, and I'm told a lot of that is done by hp for the hp-ux versions. It probably costs a massive amount more to try and make the base code for CMT, which is very different to classic EPIC/RISC CPU code. If it was merely an economic argument then Larry would have stopped development on CMT and sold what little was left of the Sun hardware carcass to Fudgeitso. Maybe this is an attempt to make hp buy the Sun hardware bizz, like Oracle wanted hp to do in the first place, or maybe Larry is trying to force hp into supporting Slowaris on Itanium. I'm told by the hp and IBM reps that they had a big target for the year - all the customers still on SPARC-Slowaris - but it looks like Larry has just given the IBM and hp sales teams a new set of opportunities - all those accounts using Oracle that can be sold DB2 or M$ SQL instead!
By your logic, Oracle should have dropped SPARC, not Itanium. Itanium passed SPARC in revenue and units shipped in 2008 or so..
Larry is insane
This seems suspiciously like a case of "We can't make a compelling hardware platform, so lets hamstring those who can." It may seem like a good idea in the short term, but I suspect most of the customers who are going to be forced to migrate because of this will be extremely wary of doing so to Oracle hardware - DB2 may prove to be the real winner here. If IBM were smart, they'd re-energize Informix and/or DB2 for HP-UX and start trying to poach Oracle customers that don't want to switch OS.
Kiralexi - Wrong way round..
No one care's about what OS is running, that's the whole point here!!!
App's rule the roost, what hardware and OS is driving them is irrelevent. Oracle customers will switch OS/hardware far, far faster than they switch database.....Larry knows this, hence him playing this card.
I work for one of the biggest yank banks going, the Oracle DB is not going anywhere (even our Sybase is phasing out to be replaced by Oracle), whereas Solaris and AIX is going out of the door for the low end DB's and middleware to be replaced by Red Hat.
Our M9000's running our biggest DB's will eventually be replaced....by another big SPARC box, why? Intel cannot scale x86 that big. The day it does is the day we will buy it. What chance POWER did have to replace the big M-series box's is now dead, as Oracle will simply price their hardware lower than (the admittedly better) P-Series to keep us on SPARC, and they will be doing this for every account going.
RE: Kiralexi - Wrong way round..
".....Oracle will simply price their hardware lower than (the admittedly better) P-Series to keep us on SPARC....." Yeah, and that strategy worked so well for Sun! Oops! - no it didn't, even trying to flog their kit 20% lower than IBM or hp still didn't save Sun. Walking around with their pants round their ankles just made Sun fall over all the faster.
Take my company as a case study, we're the enterprise type of bizz Larry will want to sell SPARC to, and we run a lot of Oracle on Integrity and Power. Consider the options:
1. Stay on hp-ux on Itanium and find another DB (we don't really use much else from Oracle), but then we'd have to revalidate the app stack. Painful, messy, and risky to data, but we do have DB2 tested and ready to go. I'm grinding my teeth at that idea, but it's bearable, the main risk is in getting the data out of Oracle and into DB2. A loser for Oracle.
2. Stay with Oracle DB, just move to another server platform. Not as painful as we already have IBM Power scoped as a backup choice, but no win for Oracle there then unless they also bump up license costs on Power, and they just make the board twitchy about trusting Oracle.
3. Larry's preferred option - replace all the hp-ux kit with Slowaris on SPARC. Not even on the table, for so many reasons, but mainly because Power is such a better option.
4. Switch to x64 where we can. Very painful, more so than the switch to Power. Larry still wouldn't win anything here, in fact he would probably lose more as going to x64 leads to all kinds of unfavourable comparisons with M$ SQL and/or FOSS alternatives. The x64 option gives far more choices for moving away from Oracle, not towards them.
All in all, this strategy seems to benefit IBM more than Oracle. Larry's not stupid, at least not THAT stupid, so I'm guessing there must be more going on here.
SP&L or whatever....
Well well, Larry it taking his love of the art of war a bit seriously isn't he!!
Matt B - I'd love to hear your explanation for this one...
I love it though, how often do you see naked aggression of this kind in the enterprise market? Never. Like I said before, Itanium was dead when Red Hat and MS dropped support, now that the dominant DB vendor has also dropped it, HP-PHUX is finished...no if's and no buts.
Oracle is essentially abusing it's monopoly position in the DB market, it wants to dominate the high end, high margin space, where it's only remaining competition is now IBM. I suspect Oracle will alter AIX pricing again to favour Solaris, and push as many people towards SPARC as possible.
It's downright nasty, no doubt about that, but that's business and Larry is bloody good at it...
Perhaps... he went to far this time our old frind larry
Well perhaps he went to far this time.
Had a client meeting here the other day with one of our big clients, and funny enough the balance had changed. The Oracle DBA's who normally get to do pretty much what they want. And they aren't that bright.. sorry.. "We need RAM disks to place tablespaces for sorting on, disk is slow so we need RAM disks' .. gee dorks.
But this time the tables had turned, it was actually the Infrastructure architects (an old UNIX guys) who had the final say. And the message was quite clear, help us dear outsoucing partner, how do we pay less in Oracle licenses.
Now I had a field day :)=
RE: SP&L or whatever....
I'm guesing the whole "split-brain" thing included a large amount of memory loss? For those with a short memory, go back and look at Larry's flip-flops over RedHat - first he loved them, then he hated them, then they partnered with them, then they did an attack campaign on them, etc, etc. It's almost a shame that hp-ux doesn't have a monkey as a mascot like Tux the Linux penguin, as I think Larry would look just right in a monkey suit for when he is forced to flip back behind hp-ux!
RE: Perhaps... he went to far this time our old frind larry
I'm dusting off the report on the last POC we did with DB2. Just been asked to run through it again. Damn you, Larry, I was looking forward to a quiet Summer!
Good thing HP and others make good value x86 servers
I think Oracle have killed itanium support not just to save money, but to stop people buying Itanium+Oracle for high end/High Availability servers. Now you will have to buy Sparc for that, but I don't personally see sparc as that high end (I could be wrong). It helps oracles margins, and shows the danger in keeping your data in Oracle. They don't just get your database money, they now get to say which OS and servers you buy.
Oracle: like the old IBM. And making lots of enemies along the way.
The whole story about abandoning Oracle SW for Itanium is just Mark Hurds revenge.
So you have to read the sentence "we're not going to support Itanium any longer" as
"Larry, I want to spit again in hp's face."
Remember: since Mark Hurd joined Oracle and got responsible for ~2/3 of the company, specifically controlling the whole hardware zoo, SPARC server revenues dumped heavily.
Loosing about 20 to 25% revenue, IIRC, compared to the quarter before.
So Mark _has_ to do something to help SPARC based systems ramp up again.
And, yes, he still hates hp..
HP BCS and Converged Infrastructure Hunter reps
IBM Power Systems is hiring for not only a growth platform but with a better % of revenue plan than HP. You also get reimbursed for mileage vs. having to pay money to drive a boring company car.
ibm.com/jobs click on Search put in hunter
- Product round-up Too 4K-ing expensive? Five full HD laptops for work and play
- Review We have a winner! Fresh Linux Mint 17.1 – hands down the best
- Vid Antarctic ice THICKER than first feared – penguin-bot boffins
- 'Regin': The 'New Stuxnet' spook-grade SOFTWARE WEAPON described
- You stupid BRICK! PCs running Avast AV can't handle Windows fixes