Ever since news got out that Oracle snapped up Sun Microsystems, a big question mark has been hanging over Sun’s SPARC business. Would Oracle kill it off? Sell it? Or would they actually invest in creating a top-to-bottom hardware+software stack, like Apple and Cisco? Oracle’s CEO has given the answer. And yes, that’s a new category icon up there.
In a SEC filing yesterday, as well as during a talk with Reuters, Oracle CEO Larry Ellison explained he wants Oracle to become like Cisco and Apple. “We’re definitely not going to exit the hardware business,” Ellison said, “[Apple and Cisco] enjoy very high-margins because they do a good job of designing their hardware and software to work together.”
He also said that Oracle will stick to the SPARC architecture, which can only be seen as a win for us enthusiasts. “Once we own Sun we’re going to increase the investment in SPARC. We think designing our own chips is very, very important,” Ellison said, “Right now, SPARC chips do some things better than Intel chips and vice-versa. For example, SPARC is much more energy efficient than Intel while delivering the same performance on a per socket basis. SPARC machines are much less expensive to run than
Intel machines.”
He further explained:
We want to work with Fujitsu to design advanced features into the SPARC microprocessor aimed at improving Oracle database performance. In my opinion, this will enable SPARC Solaris open-system mainframes and servers to challenge IBM’s dominance in the data center. Sun was very successful for a very long time selling computer systems based on the SPARC chip and the Solaris operating system. Now, with the added power of integrated Oracle software, we think they can be again.
Even though these are just words, they should reassure SPARC enthusiasts and customers. Only time will tell, however, how much these words are worth.
Larry Ellison is hardly going to say “Yer, we’re going to savage 20,000 jobs”. Keep in mind that this is not a done deal yet and he has some sweet talking left to do.
One thing that I’ve learned is that Oracle are ruthless when it comes to costs and returns. They’re not a great technology company by any stretch of the imagination and even worse at creatiing a ‘brand’, so that whole thing about being like Apple and doing hardware is hilarious (What did happen to that ‘Network Computer'(tm) Larry?), but their attention to costs and returns has given them a cash pile that has allowed them to buy Sun and ride the current storm. Honestly, Sun need some of that nous desperately.
SPARC isn’t going to just fall off a cliff, but one thing I could see coming is selling it back to Fujitsu completely – if they’ll have it – and cutting the costs of developing the platform. They’ll still sell SPARC and use it, which might be Larry’s sleight of hand, but it won’t be backed by Sun with development money. I don’t know whether Fujitsu could make it pay though. Sun need to cut their exhorbitant costs here versus their declining margins, something that they have failed to do over the years.
True to form, Oracle will look at the bottom line with server sales versus development costs:
http://www.theregister.co.uk/2009/04/28/sun_fq32009_numbers/
That doesn’t paint a pretty picture. Everybody’s sales are down, but when you’re reliant on server sales as much as Sun are then it’s difficult for Oracle to keep justifying SPARC unless they can make the margins far bigger. You never know though. Oracle have been known to sell sand to Arabs and it’s what they’re good at. The only reason IBM makes money out of Power is because of what they sell with it. Server sales themselves are not enough.
Uhhhm… Oracle RDBMS? Admittedly not as advanced as Sqlite or MySQL. But it shows some promise.
Uhhhm… “Oracle(tm)”? Admittedly not a household name like, say, Firebird DB. But has gained a certain degree of mindshare in hardcore IT circles. [/q]
But one thing that pretty much everyone can take for granted is that whenever there is a story about Sun or *Solaris, Segedunum will be quick to spew his vitriolic “thoughts” on the matter.
While I’m pretty much pro-anything-posix, I’m hardly a Sun fanboy; I’m more focused on Linux. But even I am weary of the predictable venom he spews at Sun and *Solaris. Or is it more generally at anything non-Linux?
Edited 2009-05-08 15:28 UTC
Ummmmmm, yer, and? I once went on an Oracle training course years ago where even the tutor was poking fun at some of the things that Oracle fixes in successive versions so rolling support fees can be accumulated. It’s great if they can do that but I think anyone worth their salt should know that Oracle’s database systems are not the greatest software you’ve ever seen, and particularly their management tools that lurch around.
Ummmmm, the reason why that comment was made was because of the whole Apple tac that Larry hinted at. I’m not entirely sure why you’re comparing Apple as a brand to Oracle and Firebird nor why you think you’re being clever by sticking ™ after Oracle. No really, you don’t know what you’re talking about.
One thing I do take for granted around here is the complete total and utter lack of discussion on anything approaching a rational level. If someone sees something they don’t like they have a go at the person making the comment, with little to no discussion in return, and if they cannot put something down with that that then they go utterly mental. I expect to see the usual suspects in short order ;-).
Blah, blah, blah. I choose not to be politically correct because that just makes discussion pointless. I’m afraid the execs and the politics at Sun took what could have been a company pulling in billions in profit right now and flushed it. I don’t make any apologies for pointing it out.
Sorry mate. You obviously didn’t see my comments on the Mark Shuttleworth and Wine article ;-).
If you have something useful to say by all means say it. Alas, we’re already umpteen miles off-topic in the first comment. C’est la vie. Par for the course.
Oracle is definitely a brand; most of my not techie friend and relatives recognize the name and know that it’s a database. ( Yes, I know that should be DBMS, but the average man-in-the-street doesn’t)
Then your friends aren’t as non-techie as they think they are. The very fact that they have just associated Oracle with ‘being a database’ (probably at some organisation they work at) shows that Oracle has a very steep climb in knowing what brand they want to create and being attached to it in the future.
Keep this in mind when you read below.
This coming from you is incomprehensibly hilarious.
Remember what I asked you to keep in mind above? That is a prime example the Pot-Kettle cliche that you so shamelessly embody.
Huh? Sell “SPARC” to Fujitsu? a) SUN does not own “SPARC,” and b) Fujitsu is actually trying to unload some divisions, among them their SPARC team.
Oracle wants to control the whole stack, from HW, OS, SW, and services. Like IBM, and giving them an edge over Microsoft who don’t control their HW stack. Niagara is a pretty good performer in the space Oracle is interested in, so I doubt they will be canceling it any time soon. BTW, over 75% of SUN revenues come from HW (and a big chunk of them are from SPARC-based systems). Why the #$@# would Oracle buy SUN if they were not interested in their HW?
The hint that I made was that Sun would sell it’s current and remaining SPARC-based and hardware development to Fujitsu and completely outsource it. For anyone who has a firm grasp of English and isn’t dense then they should be able to work that one out.
Errrr, yer, which is why I wrote this:
If Fujitsu is having trouble, who else can make it work? That’s why Sun outsourced much SPARC development to Fujitsu in the first place.
Nice one Sherlock.
If you, and even your partner like Fujitsu, cannot make a hardware platform pay, your costs are too high and you’re shipping less and less of it over time then do the maths. $309 million in billings is just not on for a platform you are investing a lot in.
Which is why I said they rely a lot on hardware and which they’re selling less and less of each quarter meaning that the margins they have to make off them to make it viable are astronomical.
Their customers. That’s the only reason IBM was interested. They also get a toe-in in a lot of markets where they have’t been before. Oracle aren’t particularly interested in whether they sell SPARC or x86 to customers as long as they can sell a lot of them at high margins with lots of add-ons and they’re growing. Do the maths.
I think we’ll be tripping the light fantastic on this though.
Which customers? The ones that wanted to buy expensive servers? What will you sell them if you aren’t selling them expensive servers? IBM would have probably killed Sun Sparc to get the new customers to by POWER. It doesn’t make sense for them to pay for customers that they would then just abandon.
If they did buy them for the software or the software customers, Solaris on x86 & JAVA then it makes sense to kill the SPARC hardware.
Blah, blah, blah.
Since you are so quick at criticizing other people’s reading and comprehension (we’re supposed to be able to read your mind it seems). Riddle me this: Why would SUN outsource their SPARC development over to Fujitsu, when Fujitsu themselves are trying to outsource their own SPARC team?
Also, right now Niagara is a pretty nice Cash cow for SUN. Again, why would they outsource/kill it/stop developing it? Esp. since Niagara 3 is pretty much on schedule, it is brining plenty of revenue, and it is one of the top performing platforms for ORACLE?
What? Seriously. What? We’ve been over this, and you’ve inadvertently answered your own question for me and solved this on your own, so it is a riddle bizarrely. You’ve even begun to work this out by yourself, so well done.
The reason why Sun would outsource SPARC development completely is that their development costs are too high versus their returns on it, their sales and shipments have declined still further and unless they get sold then they’re ultimately going out of business. Sun are in serious trouble in case you hand’t noticed.
If Fujitsu can’t make SPARC hardware development work as a business either, with them being better suited to hardware development, then we are left with one conclusion – SPARC as a platform isn’t viable from a business perspective.
No it isn’t. Saying it won’t make it true. Look at the cold hard billing figures. $309 million. Then look at the development costs involved in keeping hardware rolling, developed and maintained. It’s pathetic to keep a platform going with those returns.
That is $309 million a quarter. That too a quarter where the IBM takeover rumor probably put a lot of customers on hold from completing orders.
That is more than $1.2 billion a year in revenue. If you look at the figures the product line has been growing year over year triple and double digits. That is one product line in Sun’s portfolio. You kept claiming the Niagara product line was only making $100 million and not growing. I can pull up past threads with your stupid posts. The data clearly shows you have no idea what your are talking about.
Do you have any cold hard facts proving the product isn’t making its margins?
To put things in perspective. VMWare’s yearly revenue is $1.9 billion. Salesforce.com makes $1.0 billion a year in total revenue. Maybe they should all close up shop.
Edited 2009-05-10 21:04 UTC
Maybe you should be taking English lessons because if you actually read the QA…
> No, we are definitely not going to exit the hardware
business. While most hardware businesses are low-margin,
companies like Apple and Cisco enjoy very high-margins because
they do a good job of designing their hardware and software to
work together. If a company designs both hardware and software,
it can build much better systems than if they only design the
software. That’s why Apple’s iPhone is so much better than
Microsoft phones.
Like the poster said Oracle wants to control the entire stack and if you had paid attention to the original purchase this would have been a nobrainer from the perspective of Oracle.
What Oracle is realizing as has IBM, as has Apple and as is Cisco is that they want to sell boxes with software.
Oracle wants to sell a turnkey solution to its clients.
I actually think Oracle will invest in Sun, Sparc, SunOS and get rid of everything else. Oracle got an entire processor design system for peanuts!
They like Java because it allows to sell boxes where clients can deploy their own business logic using toolkits that already exist.
It is actually pretty brilliant strategy.
Whereas other companies like Microsoft have to create a complete stack, figure out the performance, etc, etc. Oracle and IBM just sell a turnkey box.
I was talking yesterday to an Oracle and SQL-Server DBA. And he said Oracle on Windows is ok in performance. But the moment you move it to a UNIX box Oracle becomes IMPRESSIVE…
it’s to soon to tell how this will turn out, but if ROCK doesn’t save SPARC then it’s over.
That’s pretty much it, and you’ve said it better than I have.
ROCK is pretty much DOA, it is 2 years late and like Millenium (US V) they will most likely can it.
So when something (the T1/T2) grows at 80% per year – that’s apparently DOA. DOA means, Dead on Arrival – if it were dead then sales would have failed to take off, and there would have been no growth. You also ignore SPARC64 VII which Sun is now using from Fujitsu, and the steps to reduce duplication and overhead associated with two competing implementations of the SPARC architecture over a small market (dollars and units shipped).
Regarding the delay of their next version, Rock – from what I understand it will be pin compatible with the socket used in T1/T2 and when you upgrade from T2/T1, there is a discount (as they have done in the past). So I can’t work out why you’re having to resort to scaremongering and half baked truths simply to push the agenda that you hate SPARC, Solaris, Java and Sun Microsofts/Oracle.
With Oracle owning Solaris, they can scrap their Linux project. I think this affects Red Hat the most, since Sun was trying to replace Fedora distributions with Solaris. Now Oracle has the muscle to make that happen.
this comes across as pretty ridiculous. Everything is being commodotised. Oracle making new investments in sparc seem pretty ridiculous. Not even rock could save sparc.
Apart from IBM, everyone else has gone on the x86 bandwagon. Maybe intel will pull a rabbit out with the newer Itaniums…
I dunno, I just dont see any future for sparc at oracle. high end chip dev is now a waste of money.
Do you even know what commodity is? do you realise that x86 isn’t an open standard where as SPARC is? do you realise that OpenBoot is an open specification? do you realise that everything within a SPARC system is open standards based? You ramble on about commoditised hardware and yet you’re talk about the Itanium.
Pardon – but you need to explain the point of your post because so far it has absolutely none – apart from spreading fear, uncertainty and doubt (FUD).
do you know what commodity is? just because ofw is a ieee standard means nothing. just because sun released the schematics for sparc does not mean its commodity. where are all the off the shelf fabbed sparc chips by everyone but sun? (fujitsu sparc is totally unrelated to suns openly released design).
whats fud about saying sparc is not a commoditised platform?
SPARC customers can always roll their own chips. After all the whole thing is GPLed, meaning that if you can find an FPGA big enough to fit it, you’re welcome to advance the SPARC platform in any way you see fit
Also, to the person who said “everyone is jumping on the Intel bandwagon”, where do you get that from? There’s a certain architecture called ARM which would beg to differ with you. No it’s not in servers (yet) but it’s crtainly doing well in the market.
Edited 2009-05-08 19:59 UTC
x86 doesn’t compete with ARM because nobody has really built any easily embeddable x86 chips for use in cameras, cell phones, or mp3 players. Intel and AMD are both trying to aim for those platforms, and I think there’s a fairly good chance of x86 dominance.
After all, x86 took over the desktop away from PowerPC (and m68k before that). They stole the workstation market from PowerPC, MIPS, SPARC, PA-RISC, Alpha, and others. They stole the server markets from the same chips.
The blade server was invented with x86, as was the netbook. Those are new markets that didn’t exist with other chips.
x86 is definitely penetrating the HPC market.
The mainframe market hasn’t really changed much in 20+ years. That one is still safe.
that was me that said that, and uh, you do realise we are talking bigiron, hpc, massive cpu clusters like the sun fire e25k, 15k, etc and such, not wristwatches. arm is doing great in cellphones but we are not talking about cellphones.
These nice words from Oracle are just pre merger/takeover niceties to keep the Sun people onside until the shareholders approve the deal.
Then…. Wait a few days and the Oracle Axe will fall on many Sun departments. In the following debacle, many key people will vote with their feet and get out of Oracle before they are pushed.
Two people I know at Sun have already done this. They are Java Middleware developers and already the signs for a complete team clearout are on the walls with senior management keeping their heads down in their cubes.
Wang, Apollo, DEC and now Sun. The same thing will happen that happened to the others. They will own some new trademarks, save what could profit them and chop up the rest into little bits and flush them down the toilet.
I did find someone’s comment about the REDHAT/Solaris competition interesting…
I see the new Oracle logo has been replaced by the old Sun one……….. 🙂