Oracle to stop all software development on the Itanium processor

[taltamir]

I'd make a bet on it but its a shame neither of us will remember this conversation in 10 years .

 

[Voo]

I'd make a bet on it but its a shame neither of us will remember this conversation in 10 years .

Yeah I thought the same thing, too bad

But I made an entry in my google calendar anyhow, so if we're both around in ten years, don't be surprised if you get a PM.

 

[taltamir]

That would be awesome.

 

[Idontcare]

HP Calls Oracle Move ‘Shameless Gambit’ to Hurt Competition

Oracle Corp.’s plan to drop support for a server chip made by Intel Corp. is a “shameless gambit” that jeopardizes customers and will cost hundreds of millions of dollars in lost productivity, Hewlett-Packard Co. said.

Oracle said yesterday that it will stop developing software for Itanium-based servers because the processor is “nearing the end of its life.” Intel denied that assertion, saying it will continue to produce Itanium and has plans for future versions.

The move rankled Hewlett-Packard, the biggest producer of server computers that use Itanium, in part because Oracle now competes in the server market. Through the acquisition of Sun Microsystems last year, Oracle gained a lineup of servers that run an Itanium rival called Sparc. By pulling the plug on Itanium software support, Oracle is undermining competition in the market, Hewlett-Packard said today in a statement.

http://www.businessweek.com/news/20...ove-shameless-gambit-to-hurt-competition.html

Not a surprising position by HP

 

[taltamir]

Not a surprising position by HP

care to elaborate? I am curious why you think that

 

[Idontcare]

care to elaborate? I am curious why you think that

They've obviously got to initiate damage-control after Oracle's FUD statements made yesterday.

 

[bryanW1995]

it's good to see that oracle and HP have put their differences (cough, mark hurd, cough) behind them...

 

[Idontcare]

It is an incestuous industry.

 

[Idontcare]

Hmmm...this gets a little more interesting with each passing press release. You got to admit there are some good points made here.

http://www.brightsideofnews.com/new...2c-starts-war-of-words-with-intel-and-hp.aspx

"Just the opposite is true," Oracle said with its latest volley in the Itanium skirmish. "Oracle has an obligation to give our customers adequate advanced notice when Oracle discontinues development on any software product or hardware platform so our customers have the information they need to plan and manage their businesses." "HP is well aware that Intel's future direction is focused on X86 and that plans to replace Itanium with X86 are already in place. HP is knowingly withholding this information from our joint Itanium customers."

 

[taltamir]

ask yourself, would you feel comfortable developing for itanium? do you believe that intel is really comitted to itanium? or does it seem like a dead end platform and waste of your development efforts?
I can hardly see how oracle can be accused of maliciousness here. As much as I hate oracle and as malicious as they are; they are absolutely in the right in this case. Their development efforts are not a charity and there is no reason for them to work on intel's abandonware.

 

[Idontcare]

ask yourself, would you feel comfortable developing for itanium? do you believe that intel is really comitted to itanium? or does it seem like a dead end platform and waste of your development efforts?
I can hardly see how oracle can be accused of maliciousness here. As much as I hate oracle and as malicious as they are; they are absolutely in the right in this case. Their development efforts are not a charity and there is no reason for them to work on intel's abandonware.

Would you expect IBM to work on developing software for Itanium?

What about SUN before Oracle bought them?

 

[pm]

ask yourself, would you feel comfortable developing for itanium? do you believe that intel is really comitted to itanium? or does it seem like a dead end platform and waste of your development efforts?
I can hardly see how oracle can be accused of maliciousness here. As much as I hate oracle and as malicious as they are; they are absolutely in the right in this case. Their development efforts are not a charity and there is no reason for them to work on intel's abandonware.

Why do you call it "abandonware"?

 

[podspi]

FUD in its truest form. I'm a little surprised Oracle wants to push SPARC, given its open nature. It really is a shame Sun wasn't viable on its own. They truly gave the world such great technology, and asked for too little in return.

 

[Martimus]

FUD in its truest form. I'm a little surprised Oracle wants to push SPARC, given its open nature. It really is a shame Sun wasn't viable on its own. They truly gave the world such great technology, and asked for too little in return.

Solaris was extremely expensive. We had a ultra sparc 2 workstation that was given to my college roommate about 10 years ago. It was interesting, but the liscenses on the software on it was about $10K if I remember correctly.

 

[Idontcare]

Solaris was extremely expensive. We had a ultra sparc 2 workstation that was given to my college roommate about 10 years ago. It was interesting, but the liscenses on the software on it was about $10K if I remember correctly.

The money to be made in the big-iron market is almost entirely from software licenses and support contracts. Hardware revenue is usually a break-even situation.

 

[podspi]

The money to be made in the big-iron market is almost entirely from software licenses and support contracts. Hardware revenue is usually a break-even situation.

Ah. Good old vendor lock-in.

 

[Edrick]

Anyone else have a feeling that Poulson could change a lot of things if it performs as expected? Saviour of the Itanium maybe?

 

[taltamir]

Why do you call it "abandonware"?

alright, abandonware is going too far. Intel still updates it, just not nearly with as much gusto as x86

 

[bryanW1995]

FUD in its truest form. I'm a little surprised Oracle wants to push SPARC, given its open nature. It really is a shame Sun wasn't viable on its own. They truly gave the world such great technology, and asked for too little in return.

don't worry, that's about to change!

 

[bryanW1995]

ask yourself, would you feel comfortable developing for itanium? do you believe that intel is really comitted to itanium? or does it seem like a dead end platform and waste of your development efforts?
I can hardly see how oracle can be accused of maliciousness here. As much as I hate oracle and as malicious as they are; they are absolutely in the right in this case. Their development efforts are not a charity and there is no reason for them to work on intel's abandonware.

the problem with this announcement is that it is clearly in oracle's best interest for itanium to just "go away", and if oracle stops writing software for it they might just be able to hasten its demise. If not, then if nothing else they could at least sell more sun-based systems. This is a classic case of IBM-style FUD like that used against amdahl in the 70's. Did somebody at oracle hear/assume something? probably, and intel/hp certainly aren't going to say "don't buy itanium b/c we're stopping it in 2 yrs", but all in all it's clear that this would not have been announced if 1. oracle hadn't bought sun, and 2. oracle/sun was kicking butt/taking names with sparc sales.

alright, abandonware is going too far. Intel still updates it, just not nearly with as much gusto as x86

of course, x86 is a much greater % of their business/profits, it makes sense that they would focus more time/effort on that than itanium. however, a small % of intel is greater than 100% of most companies.

 

[Voo]

the problem with this announcement is that it is clearly in oracle's best interest for itanium to just "go away"

Sure SPARC played some part in their decision, but then Intel REALLY is the last one who should complain about anticompetive behavior against the competition

And then, there's the question.. what architectures were MS and Redhat selling after they stopped supporting Itanium months ago? So how much are the few % of Itaniums in the high-end server market worth investing in? Microsoft, Redhat and Oracle (and all the other software producers who never sold Itanium builds) think not that much and at least two of the three are more or less neutral in that discussion.

 

[taltamir]

the problem with this announcement is that it is clearly in oracle's best interest for itanium to just "go away", and if oracle stops writing software for it they might just be able to hasten its demise. If not, then if nothing else they could at least sell more sun-based systems. This is a classic case of IBM-style FUD like that used against amdahl in the 70's. Did somebody at oracle hear/assume something? probably, and intel/hp certainly aren't going to say "don't buy itanium b/c we're stopping it in 2 yrs", but all in all it's clear that this would not have been announced if 1. oracle hadn't bought sun, and 2. oracle/sun was kicking butt/taking names with sparc sales.

1. They are still developing for x86.
2. Intel is actively blocking nVidia from selling compatible hardware and wants to complain that their competitor is not longer willing to actively prop up intel's unloved stepchild?
3. It hadn't occurred to me, at the time of posting, that Oracle owns a competing platform to itanium... No wonder SUN failed commercially.

 

[theeedude]

Intel should just add RAS features to their Xeons and ditch Itanium for good.
It's a failed experiment that should have been ended when it became clear that it wasn't going to take over from x86. If the point was to do a niche server architecture, HP had both PA RISC and Alpha to choose from at the time.

 

[Edrick]

Intel should just add RAS features to their Xeons and ditch Itanium for good.
It's a failed experiment that should have been ended when it became clear that it wasn't going to take over from x86. If the point was to do a niche server architecture, HP had both PA RISC and Alpha to choose from at the time.

- Intel is already adding RAS features to the Xeon EX lines.
- Itanium is a better uarch then x86. It just never got the support.
- Itanium replaced PA RISC and it borrowed a lot from Alpha (same engineers)

 

[theeedude]

- Intel is already adding RAS features to the Xeon EX lines.

Good, then it can replace Itanium.

- Itanium is a better uarch then x86. It just never got the support.

That's arguable. EPIC is very compiler dependent to extract parallelism and schedule instructions to specific functional units, something x86 and other architectures do in hardware. When you do it in compiler, you have to evolve your architecture tightly coupled with the compiler, which slows down progress. So whatever advantages Itanium had from the beginning, either never materialized due to insufficiently good compilers, or x86 ran laps around it through iteration.
Also, Alpha had Windows NT support, they should have used that if they were going to switch architectures. At that time, Alpha was already designing a chip with "hyperthreading" (SMT).

- Itanium replaced PA RISC and it borrowed a lot from Alpha (same engineers)

It's a completely different VLIW architecture as opposed to Superscalar for Alpha.
It's mostly HP PA RISC engineers in Fort Collins that got borrowed. Alpha guys in Mass had many more options with the .com boom starting up.