Oracle Won't Abandon SPARC, Says Ellison 280
fm6 writes "When the Oracle acquisition of Sun Microsystems was announced, it was widely assumed that Oracle was interested only in Sun's software technology, and would sell or discontinue all its hardware businesses. Larry Ellison, in an interview just posted on the Oracle web site, says that's not what's going to happen. In particular, SPARC isn't going anywhere (PDF): 'Once we own Sun we're going to increase the investment in SPARC. We think designing our own chips is very, very important. Even Apple is designing its own chips these days.'"
Are You Really Prepared for the Hardware Market? (Score:5, Interesting)
My theory why: multiprocessors (Score:5, Interesting)
So, just dump more processors in a box, and optimise the processor's design to your needs.
Apple figured it out, and Oracle's not stupid. This should work until the next big jump in processor design.
RS
Re:Of course (Score:5, Interesting)
Servers were Sun's highest margin stuff? No wonder they plummeted and got bought. But if Oracle doesn't find value in offering servers bundled with software, one would wonder why IBM does. It's pretty clear that servers are now second fiddle to IBM's software business.
Is it just me or was he explicit about maintaining Sparc, but said nothing about x86 servers? I'll have to find the rest of the interview on Reuters.
Re:Of course (Score:4, Interesting)
For as long as I can remember, Apple has been designing and outsourcing their own chips. Be it in the form of custom ROMs or VLSIs wich Apple is a big user of.
Sort of a weird line coming from my boss whose also on Apple's board of directors.
What I think he meant was to emphasize that while Apple uses Intel and makes it's software (like Oracle), they also design their own chips (more so where the AIM alliance's desktop grade PPC was viable).
Re:Are You Really Prepared for the Hardware Market (Score:5, Interesting)
Oracle, on the other hand, is extremely good and marketing. They are especially good at marketing to business. They are also good at knowing what businesses want (or alternately, making business people want what they have). I don't like Oracle, but I have to say this may be the best thing that's happened for Sparc in a long time.
Re:Good for routers? (Score:4, Interesting)
Anyone betting on Microsoft buying out AMD? (Score:5, Interesting)
I'll wager this character "!"
I have a funny remark regarding what Ballmer is doing but my post would be tagged as flamebait, so I'll just write the clencher: Toilet paper.
More than routers (Score:5, Interesting)
It's not routers -- it's specialty appliances.
Take for instance your GPU -- it's just a processor that's tuned to do one specific task. Now, imagine that Oracle could take Sun's experience to customize a chip for the type of instructions that their database used a lot. Sure, the chip might not compete on all tasks, but if they could give a simple drop-in oracle appliance (or even a mysql appliance, and make money by selling hardware and support for it), they might have a reason to stay in the hardware business.
Now, I don't think that they should actually make the chips -- just design them for the right balance of power consumption / integer performance / floating point / cache / whatever makes sense for their applications.
Oh -- and to answer your question -- Sun is Oracle's recommended software platform. And Sun bought the Cray assets from SGI -- the E10k and other 5 digit models are descendants of that line. SPARC are highly reliable, high performance processors (or at least, they were back when I used to work on Suns ... from 1995-2003) -- but it's like RAID -- if you can throw 10 cheaper processors at it, do we really need the one big one? And that all depends on what you're trying to run on it.
Re:Change in the wind.... (Score:4, Interesting)
Can you amplify on this? I tried an intel-based netbook recently and was pretty dismayed at the performance. I have an ARM-based music server running Linux, and although it's fine for the purpose I'm using it for, it feels agonizingly slow when I ssh in and do things on the command line -- I shudder to imagine what it would be like running Gnome and OOo or Firefox on that CPU. It seems unlikely to me that anyone could make an ARM-based netbook with acceptable performance any time in the near future, unless they were using software much more lightweight than Gnome, OOo, and Firefox. And yet I hear people talking as though ARM-based netbooks will be on the market within a year or something. What am I missing here? Is it all vaporware? Are clock speeds of ARM chips improving at some fantastic rate? It's one thing to run software like iPhone apps that are designed from scratch for a low-end CPU, but I just don't see how it's going to happen with a more traditional desktop software stack.
Re:Designing chips (Score:4, Interesting)
No, but arguably starting with Mac or Lisa. It's pushing the metaphor I'll admit in the sense that you wouldn't wear a Macintosh the way you'd wear an iPod. But the appeal of the Mac and the Lisa was as much or more fashion and style as it was practical.
Re:My theory why: multiprocessors (Score:5, Interesting)
I don't know about unabated. It's been progressing, but we hit a bump, and all the sudden it was all about multicore and such rather than just continuing to double the clock speed every year or two.
Look at the palpable hump in this graph [wordpress.com].
Why abandon SPARC? (Score:5, Interesting)
Doesn't that sound like they did actually want to keep all the Sun's hardware business including SPARC from the very beginning?
You guys aren't getting it. (Score:5, Interesting)
"Oh you want support for a database product on commodity hardware? Well we have this little MySQL thing you can use.
Oh you want to continue to run Oracle? Well that is now only supported on our new line of SPARC hardware."
Oracle can now (and will) sell you the entire database from sand to sql results at whatever price they deem acceptable to themselves this quarter. You thought license costs were crazy before? Well now they come with official hardware and support contracts for the box.
Re:Are You Really Prepared for the Hardware Market (Score:5, Interesting)
I mean, how are you going to mitigate the blitzkrieg campaign IBM has launched against SPARC [slashdot.org] while you're busy with the merger details?
Interesting choice of the word blitzkrieg to characterize the marketing campaign. I think it's very appropriate.
Blitzkrieg [wikipedia.org] was a tactic to concentrate a large fast assault on the weakest part of the enemy, disregarding the flanks and trying to avoid the strong points.
It had success early on for the Germans, it was not something that could easily be maintained and after a year or so the allies were able to adapt to counter those types of attacks.
Lets not forget who won the war.
IBM is trying to take advantage of the uncertainty some people have with the merger to grab some of Sun's hardware business.
Re:Why abandon SPARC? (Score:2, Interesting)
It makes sense when you think of Apple's success in vertical integration. Why not a drop in database box that is setup specifically with Oracle's service department in mind.
Need more performance? Call up Oracle, a pre-configured plug-and-play rack mounted box arrives, you slide it in, plug it in and you have more performance.
Database Processing Unit (Score:4, Interesting)
Apple is a stealth software company (Score:3, Interesting)
Apple is a software company which happens to design hardware that they believe will run their software perfectly. It is hard to explain but, if you look at pre touch iPods, they are significantly weaker than other offerings in hardware specs. What makes people buy them is the software they run. Same thing can be said for iPhone vs. Nokia 5800. They didn't change overnight, it is same deal since first Apple 1. That is why people dreaming about official OS X on generic PC are kinda... Dreaming.
If Oracle has this neat idea of having devices, gigantic mainframe like servers (Sun's top line), portable enterprise database servers.. They are going with Apple's idea. Of course, they aren't stupid to abandon their "runs on Linux/AIX/zOS/Windows/Whatever" software.
Just imagine a Sparc which have accelerator functions just for database operations. That kind of possibilities kept Apple in PowerPC for years, G4/G5 especially have some excellent functions for media which came from Apple. Of course, times has changed and IBM started to hate end user desktop except consoles so they sold them out and moved to Intel. If you look at how easy was for Apple to move to Intel and how easy for them to release software for Windows when they want, you can't call them just a hardware company.
If something really bad happened to Apple, it is even possible to release OS X/iTools/iWork for Windows. Of course, we wouldn't get the same experience on millions of different configurations and substandard $10 cards. That is why you see Apple hardware.
Re:Of course (Score:3, Interesting)
I think many misses the fact that Sun makes some great blades running Windows enterprise... Or the entire "enterprise" market which POWER is the king, Intel, Sparc are fighting each other, AIX is a huge player, Cell started to have huge popularity as HPC newcomer etc. It is something like different universe.
Of course, I won't see a Sun workstation in my usual life, I won't sit and admin a Enterprise server but that doesn't make me treat Sparc as something so sucky that can be easily abandoned by an enterprise software company...
Come on people... Lets go to some enterprise focused sites (Register has a great section) to see the real deal before talking eh? A database company who hasn't produced any kind of "small" (sub 10 users) serving software pays billions just to kill mysql which doesn't compete in any of their segments, closes down Java, kills Sparc... If they are _that_ stupid, how come they are one of the largest software companies on Earth?
Re:My theory why: multiprocessors (Score:2, Interesting)
Well, if you want to stretch the anology (Score:5, Interesting)
Germany was a small country fighting the whole world. They lost not so much because blitzkrieg wasn't a valid tactic but because it is hard to win a battle when you fight a million soldiers and the enemy has a million in reserve.
But in this case, it is IBM who is the giant. So if you want to compare things, this is the D-day landings by the free-world/IBM vs the much beleagured Nazi's/Sun who is fighting to many battles on to many fronts and who just can't keep up with the tech race.
Analogies, you really shouldn't stretch them to far or they turn against you.
Re:Practicality and Fashion (Score:3, Interesting)
Size can't have been the thing, the first ipods were bigger than their competitors. As for the interface, well, maybe there are people who really do find that stupid wheel-thing easier to use, but there are dozens of players doing that now. So why is the ipod the one that sells? Fashion, plain and simple.
You were happy with cell phones before the iPhone came out. Most people hated their cell phones and used them only for the most basic things.
Oh, come on. Compare doing any task you can think of on a near-contemporary, say the Razr. It's barely any different, and certainly not any harder. Nope, I'm not buying it. Fashion.
Re:Practicality and Fashion (Score:3, Interesting)
Size can't have been the thing, the first ipods were bigger than their competitors.
Bigger than the low storage flash based players (flash was a lot more expensive then, a player with gigabytes of flash would have been unthinkable). But a bit smaller and a lot sleeker than things like the DAP jukebox. The UI was also pretty well designed afaict (if you are going to have a jukebox style mp3 player the interface is pretty critical).
Making a good product is all about getting things right accross the board. If your product is crap in one important area your product is crap regardless of how good it is in other areas.
Re:My theory why: multiprocessors (Score:3, Interesting)
Exactly. However, it isn't going to happen in the next few years, and from what I can gather, the next 5 are going to be the age of massive multiprocessors. There will be improvements, but nothing like the 1990s. There will be a breakthrough in speed, but Oracle is looking at the here and now, and the here and now is saying "highly specialised silicon multi-core chips arranged in a multiprocessor array", and my contention is that Oracle is responding to that reality.
We can talk about all the dreamy chips of the future, but when you have a business to run you have to look at things with definite parameters and plans over the next year or three, and that's where Sun's chip manufacturing comes in - modify the chips to fit your needs, and then dump dozens of them in a box. At the same time, optimise your code to work on these optimised boxes, and you get real performance gains in a realistic timeframe without resorting to wishful thinking of some messianic breakthrough.
Make the money you can with such a system, and when / if the breakthrough comes through, take advantage when the time comes. But basing business plans around some great technology that doesn't exist yet is sheer stupidity.
RS
Re:Database Processing Unit (Score:1, Interesting)
Let's say Oracle on specially designed Sparc does outperform other solutions. So what? I already have vastly more horsepower available than I need running PostgreSQL on commodity x86 servers. Sure, there's a market for uber-fast uber-big database apps, but the point it, that market is shrinking.
The exception might be if the uber-fast solution enables a fundamentally different approach to computing than nominally available. Maybe Oracle isn't planning so much on selling these solutions to end-users, as they are in building out a giant server cluster that they lease out a-la other so-called "cloud" computing solutions. I.e. they don't sell you sparc+oracle, they sell hosted ERP, etc. Building the servers themselves, they may be able to argue their infrastructure costs are lower than for vendors like Amazon who have to buy their servers from others.
Still, while that might buy them a little time, there's nothing preventing the creation of open hosted services. Vendor lock-in is vendor lock-in, cloud or not. Most folks I know would always rather avoid such traps; which is why there's already plenty of pressure for open standards for cloud computing services.
Re:Of course (Score:3, Interesting)
This is not so much for the Sun employees or the JAVA stockholders, this is for the Oracle DB shops that run on solaris/sparc aka the guys paying big money to Oracle now. Oracle does not want to alienate them and get them to go IBM (and then possibly DB2) or Red Hat (and the possibly Postgress/MySQL). Also now Oracle will be making revenue on those solaris/sparc shops not just for the DB. They want to at least make outward indication that they do not intend to cause trouble for them.
Re:Are You Really Prepared for the Hardware Market (Score:3, Interesting)