Predictable since 2003-ish
Jury selection is underway in the trial that pitting server maker HP against its formerly strong software partner Oracle as they argue about the future of a technology that neither one of them controls: Intel's Itanium processor. Well, given that HP has been making payments to Intel to continue to develop the Itanium chips, at …
Even if your assumptions were true - What do HP have to gain from letting Oracle get away with this?
I'd be interested to see Oracles P&L statement for oracle on Itanium - i.e. "what will it cost us to bin our most important enterprise software partner?"
Disclaimer: I'm an HP-UX lover, but don't necessarily love the ugly step-sisters - Itanium & HP Corp.
"What do HP have to gain from letting Oracle get away with this?"
The entire relationship with Oracle. HP is poking a big dog with a stick. This is just starters on how Oracle could hurt HP if they further enraged Larry. What if Oracle just said that we are not certifying or testing HP x86 or storage for Oracle software, we are not cooperating on support, HP is does not exist in Oracle's eyes? That would cause boat loads of trouble for HP. As long as Oracle is still moderately friendly with IBM, Cisco and (laugh) Dell as well as their own Sun gear, I doubt people would drop Oracle's platforms, DB, and apps in favor of vanilla HP boxes.
Agree, similar to Apple with Samsung, people buy application software, they then figure out what stack they want to run it on, they then figure out what hardware they want to run it own. As Oracle is providing the business value and HP is providing hardware that supports Oracle software, Oracle is in a much stronger position. If Oracle tells people it is either us or HP, make your decision. People will be upset and after they are done being upset they will find new hardware. The same reason HP cannot go to war with Microsoft. People will just go find some other relatively interchangeable hardware.
"....What if Oracle just said that we are not certifying or testing HP x86 or storage for Oracle software...." Well, seeing as over fifty percent of Oracle software sold goes onto hp systems of one form or another, that would be like Larry chopping both his own legs off at the knees.
Come on, people don't care about x86 servers. Oracle's software is going to trump x86 server provider every time. If Larry tells people they can't use HP anymore, they just find another x86 rebrander. No one is going to the CIO to say that they need to migrate databases much less to the CFO to say that they need to migrate their financial applications cause the system admin likes HP Insight.
".....If Larry tells people they can't use HP anymore, they just find another x86 rebrander....." If Larry did so he would end up on the end of a massive suit from hp. How exactly would he say he couldn't support one x64 vendor and not stop support for all of them, seeing as they use the same OS, chips, RAM, disks, etc? It would be a blatent attempt to disrupt hp's business and would open Oracle to being sued in the US courts.
Come on Matt, the amount of people who buys x86 and really cares about the HW is tiny. If they did care about the HW they'd buy something else .)=
Now the management interfaces and surveillance software stack that is another thing.
And it still doesn't change that.. with regards to Itanium.. then it's ... it's dead Jim. Not that I gain any pleasure from seeing another Unix platform being put on deathwatch.
I find this tricky, as I think the Itanium was a waste of time, as the world was not short of better chips than the x86 so why did they do it? Hell, HP had the DEC alpha that for many years kicked Intel's butt in performance tests...
But then again Hurd was a crap CEO who's only method at NCR, then at HP seemed to revolve around cut backs and redundancy. Oracle appointing him suggests they are arse-licking buddies or stupidly lacking in imagination.
So we have two big corporations that are hardly worth saving, other then for the memories of once-great years, and the jobs of the minions who deserve better. Just who deserves to win?
I don't understand what's the problem that HP paid Intel for Itanium development. So what? They wanted to have this chip, didn't have own resources to do it, so they paid Intel. And I assume they will keep paying and Intel will keep developing it until it will be profitable. In previous couple of years it was very profitable. It will be getting smaller but no one knows how slow/fast if not Oracle offensive move. Maybe it would be 2014 but maybe 2024, who knows. This oracle argument is taken from where the sun doesn't shine and is just lousy cover for their offensive and anticompetitive moves after idiotic Sun aquisition. I have no reasons to love or hate any of those corpos but it's very clear who's the bad guy here and who should pay whom huge penalty. In normal country anti-monopoly authoritites would take care of oracle after such move, but of course this is America, no one bothers.
I must agree Itanium is quite dead. But, I don't know how good a partner Oracle is -- since buying Sun, they've been cutting support for various UNIX systems, seemingly just to drive those customers who would buy a "big iron" UNIX system to buy a Sun from Oracle instead. I think they would have cut any x86(/x86-64) support by now if they weren't concerned about it driving away customers.
That said, this doesn't seem like something that could be sued over.
"this doesn't seem like something that could be sued over."
Wrong. Oracle has so strong position in database market that it definitely should be a target of anti monopoly investigation. I wonder why HP didn't start this yet, at least in EU. Their behaviour is clearly monpolistic and anti competitive. Software vendor shouldn't have right to use his market position to force his customers to use particular hardware.
HP is trying to push a monopoly case in the EU. Itanium is a bad case to bring forward because it is so clearly dead and Oracle can always fall back on the numerous other ISVs that dropped the platform before them. Oracle definitely is using anti-competitive tactics, but this is the regulatory system that didn't do anything about Microsoft's monopoly. MySQL was a pretty clear case of anti-competitive behavior and that did not get through.
".....Itanium is a bad case to bring forward because it is so clearly dead...." For Snoreacle's argument to hold water, it would have to explain why it is still ploughing money into the development of Oracle apps on Slowaris, which is currently making much smaller sales figures than even the depressed Itanium ones. And some of the drop in hp's Integrity server sales may be due to hp's customers waiting for the imminent release of Poulson chips (hey, if TPM can excuse a massive drop in mainframe sales with that line it applies to Integrity too!).
Software provider support is the issue with moving HP-UX to x86 at this stage. HP can port it to x86 from a technical perspective, but the ISV community doesn't want to go through the effort of creating the port and the support structure for HP-UX on x86 because the user base is relatively small to start with and the number of people who will decide to stay with HP-UX on x86 instead of just going to Linux on x86 will be really small. HP-UX, VMS, etc will stay on Itanium for the ISV support until the end of Itanium. HP should have ported HP-UX to x86 instead of Itanium years ago.
The other bitch here is that there are many sites, especially the larger HP shops, who still have a load of PA-RISC kit for many reasons.
Being asked to migrate to x86 when the move to Itanic isn't yet complete could just be the last kick that many of 'em need to get seriously pissed-off with running HP's version of the Red Queen's Race and look long and hard elsewhere for some long-term stability.
Then you'd see some really shitty numbers from HP. Much, if not most, of those Itanic sales are from PA-RISC "upgrades" for their long-suffering enterprise customers.
You're right, they should have ported HP-UX to x86 rather than Itanium. Everything you see going on now is HP desperately trying to avoid being bounced into doing just that with suicidal timing.
"I don't understand what's the problem that HP paid Intel for Itanium development. So what?"
The "so what" is that HP lied about it, big time, and so did Intel to a lesser extent.
IA64 was going to be "industry standard 64bit computing", according to Intel and the IA64 launch fanclub (including HP).
64bit x86 couldn't be done, according to Intel. Then AMD64 arrived, and suddenly Intel admitted that x86-64 was indeed possible. Intel decided it was something they had to be the best at.
Itanium didn't arrive, not in any meaningful way outside HP, and certainly not as "industry standard 64bit".
HP carried on telling anyone who would listen that Itanium was a viable chip, and that unlike the story which was used to "explain" why Alpha was cancelled, the IA64 sales revenue was claimed to be sufficient to maintain ongoing development. But that wasn't exactly true either.
You wanna buy an enterprise class system from this lot?
I read somewhere they're even talking about dropping the "Compaq" brand other than for dirt cheap entry level stuff. Talk about back to front. Compaq Presarios I've had were always cheap and nasty, HP Pavilions were fine albeit overpriced (which was fixed by buying refurb via HP Renew when it existed).
More importantly for the business computer user, Compaq Proliant has been and still is the industry standard server of choice (still outselling Dell by volume and value). x86 servers from HP? Who even knew they existed?
Do HP HQ have a clue?
Show me one chip vendor that wasn't ever lying with his roadmaps. For sure you cannot say that about either Sun/Oracle or IBM. They are as messy as Itanium. Or worse in case of sparc.
The rest of your post looks like written 10 years ago. Who cares what was going on there. There's no compaq anywhere anymore, servers are named HP Proliant since ages. It has nothing in common with current story.
The post above is correct. The issue is not that Itanium is dead (although that is also a problem for HP). The issue is that HP has been telling people that not only wasn't Itanium dead, but that other high end systems were not viable because they were not owned by Intel, did not support MS Server, etc. Sparc probably is dead, but the z/OS (mainframe) platform that was much maligned by HP will be dancing on Itanium's grave. Any customer that was convinced to migrate off of mainframe to Itanium should be furious. They were straight up lied to by HP. It is funny that Alpha was going to take down mainframe, then it was PA-RISC, then it was Itanium. Mainframe has out lived them all and going through a renaissance with Linux on z and people figuring out that IBM was right about one computer being easier and more efficient to manage than 1,000 little computers... or what is known as "cloud computing", formerly known as mainframe computing.
"Any customer that was convinced to migrate off of mainframe to Itanium should be furious"
Can't agree to that. People migrated off the mainframe cause it was plainly much cheaper. No one was so naive to believe in marketing BS about what has a future and what has not. Or rather - if anyone was so naive, it's his own problem. But I don't believe there were many cases like that. People that keep using mainframes are usually the ones who counted that apps migration/rewriting would cost them much more than savings on cheap hardware, that's all. Possibility to run linux on z is only a matter of making their bitterness a bit more sweeter, no one buys mainframe only to run linux on it.
Agree in some cases, but how costly would it have been for an enterprise customer to follow the DEC-Compaq-HP path from Alpha/PA-RISC to Itanium to whatever is next? Those Superdomes with all of HP software with Oracle on them are not cheap dates at all. Refreshing a mainframe is less costly than refreshing a large Unix system of any stripe. The people who use mainframe are the most sophisticated IT orgs in the world. If they wanted off mainframe, they would have done it over the past 20-30 years. Mainframe is still the only EAL5 secure system in the world, unparalleled uptime, unparalleled efficiency in staffing. I don't think anyone who went from mainframe to big Unix really saved a great deal of cost. They just spread the cost out among a bunch of vendors (who all hate each other now) and new staffing costs.
".....the initial "Merced" Itaniums from 2001 were co-developed with HP....." The Merced core was an all hp development, Intel's input at that point was about fabrication. Intel got into the core design development with McKinley, the first Itanium2 chip.
".....and HP-UX is the only operating system that is basically left on the processor...." Forgotten about OpenVMS and NonStop? How about Windows Server 2008 under Integrity Virtual Machines? Come on, TPM, are you getting the office junior to tech vet your work?
Haha, TPM descending to slap Matt down from upon high. In the Oracle docs Stallard from HP mentioned that MS Server of Itanium, prior to the end of life, was appx. $100 million per year which was "larger than VMS." We know that VMS is sub $100 million, how much below is unknown. VMS is really small from an install base perspective. NonStop is probably small still. Matt always forgets to mention that MS Server is done on Itanium as soon as whatever MS calls the Win8 server version is out. RHEL was done with 5. VMS and especially NonStop are massively profitable for HP, but HP-UX is the only OS with any size on Itanium.
Doesn't change the fact you said hp-ux was the only available OS on Itanium. You also forgot all the Linux variants still available for Itanium other than RHEL (such as Debian 6.0.5 and SLES 11 SP2). As to how much money hp make off OpenVMS and Nonstop I really don't know, but it seems it's still enough for the IBM fanbois to FUD them.
Most probably they don't make so much money on them (though I have no idea how much exactly) but most of customers using them are very "prestigious ones" - huge financial institutions in case of NonStop and lots of military things in case of VMS. They won't migrate anytime soon and HP will keep those up and running as long as possible.
Now that Mr Larry has decided (for some egotist, irrational reasons) to even shaft his best sales partners, this is one of the strongest arguments to avoid commercialware.
What if he contracts Alzheimer's ? Then customers will get their RDBMS only on what ? Uniys ?
Here are some free (as in freedom) alternatives:
Fine for websites and non-critical stuff, but not for enterprise DB workloads. The functional differences between Oracle's DB and those open source alternatives you could just about squeeze into the grand canyon. Also, people don't want to run a large business critical application on Ingres or one of the other flash in the pans. If the fickle open source developer community decides something else is more important next Tuesday, that project stagnates and everyone using it extensively is hosed. Do you want to bet the ranch on Ingres? Oracle acquired and put a stop to MySQL because it was their largest open source threat. If anything gets close to them again, Larry will just buy it. The developers will fragment into different forks and Oracle will keep on keeping on. If people are moving away from Oracle, it will be to DB2.
Ingres first appeared in 1974. I recall running "University Ingres" on a PDP 11/44. I'm struggling to reconcile that with the term "flash in the pan".
As regards "fickle developers", I am struck by the irony here. If I were a customer running my business on HP Itanium kit and Oracle, I could easily ask about fickle database providers. At least if I were running one of the open source alternatives, I'm not at risk of my database provider suddenly and arbitrarily deciding that my hardware isn't supported. I'm sure Oracle will be more than happy to come in and replace your kit. I'm not so sure you COO/CIO will be happy with the bill unless they have a lot more money than sense.
Yeah, Ingres technology has been around for a long time. Pretty formidable stuff too. It has not, however, been around as a coherent release or company or support organization.... They started as an IBM System R inspired DB, like Oracle, as part of BSD, spun off as an independent, did that for awhile, were acquired by CA, Postgre (Post Ingres) took over many of those users, were taken out of CA purgatory and made open source, were forked out again. If you started as an Ingres user in the 70s or 80s, you could be in several different places depending upon where the forks led you with long periods uncertainty. The current Ingres started in 2005.... Ingres is not available for commercial apps, for the large part, such as SAP and Oracle ERP. It, like the other open source DBs, is primarily for custom dev. As I mentioned, not that you can't use Ingres or Post Ingres or whatever else for development, but it isn't going to be running SAP as a full blown enterprise DB... of the sort people will be migrating from that are currently on Itanium. On Itanium, if the COO/CIO had more sense than money, they wouldn't be on Itanium in the first place.
Not that I am a great supporter of Oracle, but the above is the standard open source situation. Ingres' parent org, Actian, is venture capital owned. If they ever get large enough to really bother Larry, he will have to pay the piper like MySQL for the company. Fork it, like MariaDB, and repeat the process. I agree that Oracle is pushing people around. IBM is a viable alternative. DB2 has been consistent and nearly universally supported since way back. Basically functionally identical to Oracle, right down to PL/SQL support. If you are fed up with Oracle, DB2 is the counter weight for the large scale databases. Maybe MS SQL if the functional, clustering and scale limitations are not a problem.
Compared to AIX and Linux, HP/UX was not changed much since the late 80's and behaves as it served as template for the "Unix haters handbook". Features like an easy system management interface like smitty are nowhere to be found on HP/UX, unless one installs webmin, which works quite well. The easy in which AIX does disk management since the early 90's, is still a HP/UX (and Solaris) sysadmin nightmare.
Current Itanium hardware shows the performance specs of Xeon servers of 3-6 years ago, and is in no way close to current Power7 offerings from IBM or Xeon's on high-end system boards.
IBM managed to implement S38 and AS/400 instruction sets as microcode on Power cpu's. Strange is that HP was not able to build a Itanium microcode set for Xeon boards.
Add this up to the non-existent modernization of HP/UX and one finds the commitment of HP to HP/UX.
"....HP/UX was not changed much since the late 80's...." Sorry, did you say you haven't looked at hp-ux since the 80's? You missed all that partitioning stuff, then, that IBM's AIX still can't match (hardware-based, core-based and resource-based)? Are all the good IBM trolls on holiday?
Biting the hand that feeds IT © 1998–2019