Slashdot Mirror
Sun to Merge UltraSPARC with Fujitsu's SPARC64?
Waldmeister writes "The Register has a story from a japanese source, that Sun and Fujitsu are planning to combine their Unix server businesses. Even if Sun doesn't comment on this, they acknowledge that Scott McNealy met Fujitsu's CEO this week. If this will happen, Fujitsu will get the bigger chunk of manufacturing and engineering. With the PrimePower systems outperforming Sun's SunFire systems for some time now, this sounds reasonable, too. And it gives Sun the chance to more resources to extend their Linux and x86 business." There's also a Reuters story.
I wonder how Sun is going to get out of this long term contract with TI... otherwise, I don't see how this new merger is going to really help Sun.
Davak
... Their multi-core CPU plans.
It seems that Fujitsu is not confirming that the two companies will "broaded this relationship." See quote below.
Too much news being leaked? Or is there another reason to not confirm this at this time?
They might be doing Linux, but they are certainly not keen on the idea and are only doing so because their customers keep asking for it. Well, at least they are listening to their customers I suppose, so there is that, but it still feels to me like Sun has seriously lost its sense of direction recently. I suspect a lot of FUD filled editorial is going to be written under banners like "Has the SPARC gone out for Sun?" real soon now.
Still, at least Apple's star seems to be rising at the moment. ;)
UNIX? They're not even circumcised! Savages!
I hope this isn't the only iron in the fire fort Sun.
As others have mentioned, it will ruin their good relationship with TI.
Also, it's doubtful that special purpose RISC chips can provide enough in the price/performance arena to keep from having their market share continue to decline, as it has for the last 10 years or so.
Low end Linux servers is a dangerously competitive business for Sun to be in, but it's a growing business and one where they have much to offer.
Fortunately, if Sun "doesn't have a Linux strategy", Dell, the 800 lb gorilla, is still half-napping, too. Dell's support of Linux is weaker than that of rivals IBM and HP, plus their potentially missing some nice opportunities by actively ignoring non-Intel x86.
Sun should climb on board the AMD Opteron with Linux. They are a company with the experience and credentials to create a quality piece of hardware and have the UNIX background to cover the software side, too.
"Provided by the management for your protection."
I see two alternatives:
1. Through ignorance or oversight, you're not counting IBM, the largest computer company in the world. They maintain PowerPC (750 and 970 series and embedded stuff), Power (eServers and stuff) and mainframes (zSeries -- S/390, more than the riced out Unix that Sun wants to call a mainframe). Additionally, IBM designs ASICs not only for themselves, but for other companies.
2. You truly have an insightful point in how many people are employed by Sun-- which points to how much bloat Sun has. If, in fact, you are correct, you are pointing out how many people Sun needs to cut in order to be competive with IBM, to say nothing if Intel who only maintains a few chipset ASICs, the Pentium III/IV and the new cheap one they're advertising on TV.
My only question is who.
I could see IBM or Motorola or HP or Fujitsu as strong candidates to take over sun. Besides Sun's large bank account, they dont really have much going for them in the long run. Unless they use that money to come up with some awesome marketable product, they are done. But they have so many patents and such a large install base I just cant see them closeing shop entirely, someone is gonna pick them up. The question is who and when.
I'm sure every engineer at Sun would love to go all out and build a new RISC chip to bowl over AMD and Intel. The question is why don't/can't they?
That's what kept RISC ahead of CISC 10 years ago. SGI thought nothing of bowling over Intel's best 486 with a CPU twice as fast (R4000). That kind of performance carried a premium price. They didn't stumble until Beast and Alien got cancelled, when Intel gave SGI management cold feet.
The originally proposed quick fix is to simply redesign the the Sun servers to accept the SPARC64-V. An even better proposal, now leaked by the press, is to simply discontinue the Sun-designed servers and to sell re-branded Fujitsu designed servers. The latter proposal is a much faster path to solving the server-performance problem at Sun but leads to lower profit margins. Clearly, the situation at Sun is dire, so you can be assured that one of these proposals will be adopted. (Please read "Sun posts deeper loss for quarter". Having lower profit margins is better than having no profit margins. Right now, the second proposal appears to be winning.
Sun Microsystems will most likely fire more than 50% of its processor development team. The single biggest cause of Sun losing marketshare so rapidly is the UltraSPARC III. It has horrible performance. Check "SPEC" and "TPC".
How does this deal help Fujitsu? It can now sell more servers and get more cash. Fujitsu has the upperhand and should force Sun to accept the second proposal: Sun exits the highend server market and sells rebranded Fujitsu-designed servers. To avoid being dependent on Sun, Fujitsu should move rapidly to jettisoning the Solaris OS in favor Linux. Fujitsu is rapidly being shaped into a company like IBM: high-performance servers and computing services are the mainstays of the business.
As a side note, Fujitsu rejects hiring foreign workers (the equivalent of H-1Bs). Their SPARC64-V and SPARC64-VI were designed and built almost exclusively by native talent. When Fujitsu hires workers, Fujitsu most values the quality of "willingness to work", not "best match of skills"; Fujitsu will take the time to train its employee. Fujitsu is a traditional Japanese company that emulates most of the values that once characterized traditional American companies. Sun, by contrast, encourages the employment of H-1Bs; the UltraSPARC III and the UltraSPARC IV were built substantially by former or current H-1B workers. Sun seeks only "best match of skills" and, along with Intel, claims that they absolutely need H-1B worker even during a period of 8% unemployment among native Americans in Silicon Valley.
Where exectly did i claim that UltraSPARC was not 64 bit? Where did you get that from?
The article i mentioned shows pretty clearly that on a diagram SPECfp vs SPECint for a the common 64 bit processors there are two groups. For brevity, let's call them "fast" and "slow". Intel, AMD, PPC, Alpha, and SPARC64 are in the "fast" group, UltraSPARC and the others are "slow". Let me quote the author, Paul DeMone:
His statement may be a little blunt, but it is true that Sun servers used to outperform intel-based Linux-servers. They no longer do, and the market has noticed that. If Sun wants to sell hardware in the future, they'll have to correct that.
Please RTFABYP next time, will yer?
I agree fully. I work in the Financial industry and have seen OS/Hardware implementations stay status quo. Most exchanges have standardized on Sun and are not going to risk their reputations and financial stability on Microsoft or Linux.
Microsoft because it is not secure. Linux because there is no accountability when the $%^@ hits the fan. This however, is not my belief. I love Linux and the BSD's. Both would be good fit's in the financial world in my opinion.
I think IBM will be able to get Linux into the finacial world as their primary backend OS. But it is going to take some time convincing the big wigs and especially the lawyers.
Personally, I think Sun isn't going to go anywhere especially when it comes to their hardware. Now the OS is another story. It remains to be seen what Linux 64 Bit has in store for the future.
The question is why don't/can't they?
Sun doesn't understand OOE? They can't compete on process with Intel? They're constitutionally committed to a loser architecture? Mostly, the numbers don't add up: by the time they designed and fabbed this Intel killer -- for the sake of argument, we'll call it "Alpha" -- Intel would be ready to roll out a Pentium 8 on a process at least three generations ahead.
'jfb
To spur "enterprise Linux," Big Bang, the distributed two-phase commit.
Well, I'd argue that performance of a CPU is but a single indicator of it's overall worth.
That said, suggesting that Sun won't be able to sell hardware in the future with their SPARC processor based systems is a little short sighted and misses some key points.
Firstly, Sun is a very capable company when it comes to building systems that have more CPU's then their x86 competitors. We have 2 E4x00 machines with 12 CPU's each and they run like clockwork. Finding intel machines with many more than 8 CPU's is a challenge, and even if you do, you're crippled by the fact that the scalability is pretty much non-existent past 4 processors... especially under Windows NT. Linux isn't much better at larger CPU numbers, although I understand this is being worked on actively.
AMD has been able to correct this with the Opteron CPU's, but I have still yet to see a 12 CPU'd Opteron machine... though I'm sure someone will build one eventually. Itanium on the other hand seems to have better scalability than it's x86 grandparents... but nowhere near the almost linear performance increases seen in UltraSPARC based systems.
When you're around larger iron as much as I am, you quickly learn the value of having many relatively fast CPU's instead of having 2 or 4 really fast CPU's... with the dynamic reconfiguration ability of a good Sun machine, you could even lose 2 CPU's to a hardware failure and still have the CPU power to accommodate your customers.
bash-3.00$ uname -a
SunOS panda 5.10 Generic sun4u sparc SUNW,Ultra-2