Otellini: 'I'd rather have Sun be independent'

Well he would say that wouldn't he ! . . .

. . . having Sun be independent serves his and Intel's purpose !

Fair and interesting comment his remarks up until his point of Solaris and SPARC looking set to be EOL.

If this happens it may take a long time, SPARC and Solaris may be forced for antitrust server marketshare reasons on being spun out of the proposed merger and Fujitsu may be a most willing recipient and may wish to continue the architecture.

Also 'Polaris' isn't likely to run as fast as native even on P and Z.

Multicore and multithread SPARC appear to be well worth preserving and there may even be great value in Rock should it ever see the sunlight of day.

Much more likely is that SPARC will survive longer term, what else enterprise viable is there(?), and it may even help create a harder, better, faster, stronger and more profitable market for Solaris SPARC.

SPARC like SPARC64 are very high quality architectures now with tremendous reliability capabilities compared with early SPARCs.

Mr. Otellini's comments helps our understanding then to confirm the liekly existence of serious IBM and SUN merger talks and this would suit Southeastern Management, SAM, who's 40% shareholding is currently under water recoup some of their around USD$800M investment losses.

Likewise with the KKR's investment that preceeds even SAM.

Also the Sun board have arranged and filed covenants that help the Sun board make much money in the event of an eventual trade sale.

So it does look like due diligence and strong negotiations are happening behind the scenes maybe.

JAVA would be a grand prize for IBM, but so too would much else in Sun's massive innovative IP portfolio would be good value and great rich winnings, even if some must be let go.

Who knows, this could turn out to be the enterprise deal of the decade and maybe other players will end up kicking themselves for not having recognised superior value and acting first.

Intel included!

Stay tuned, the eventual outcome of this one is going to be very highly interesting and we're not done yet.

A shout to Tim for doing such a great job on reporting all this for Reg Readers and especially for so conveniently posting the SEC transcript of Paul Otellini's comments. Thanks Tim.

Kind regards,

Ray

@AC

Much more likely is that SPARC will survive longer term, what else enterprise viable is there(?).

At the top end - there's Power, Itanium (on a good day), but most of all I expect Intel to push x64 more that way as there development progress on Itanium has been tardy to put it mildly.

As far as other entreprise issues goes, the SPARC64 (owned by Fujistsu) is considerably better than the SUN UltraSparc processors (now effectively dead). It has much better RAS features than SUN ever put in. Whether Rock will ever appear in a real machine, who knows. SUN have canned enough SPARC chip versions already and it requires serious money to build new enterprise servers and it would leave SUN with a messy mix of legacy UltraSPARC, re-badged Fujistu SPARC64 and whatever ROCK servers would look like. All the same OS of course, but plenty of other hardware differences like the I/O systems. The space for top-end RISC is also shrinking as Intel's x64 processors eat the bottom end out of that market.

In the case of the Niagara-powered servers, they suffer from two problems Firstly single thread speed is inadequate for many applications. I've lost count of the number where a bottleneck gets hit long before the box runs out of CPU. Many critical bits of infrastructure hit that problem - T series machines work great as NetBackup media servers, but dreadfully as the master server. Any app with even one heavy single-thread process is prone to suhc problems. The other problem is, of course, x64 where you get good single thread speed, ever improving power efficiency and agrressive cost reduction (plus two suppliers, even if one is walking wounded).

The real reason for SPARC is legacy apps and the difficulty of migration. Now that IBM own Transitive expect a big push on that side should they get there hands on SUN. In 10 years time, SPARC will have gone the way of Alpha, MIPS and PA-RISC as I can't see Fujitsu managing to sustain it.

Polaris Personified?

Ray,

The Beauty/Trouble with JAVA is, is that it a Truly Dynamic Hydra Headed Application/Platform/Launch Pad with Minds of its Own. Treat it well and it Smothers you in Riches, Abuse its Novel Independent Enterprises and it will Savage you to its Advantage rather than to Mutual Benefit.

And a Perfect Clone Mirror of the Tempestuous Temptress and Madame of the Brothel/Harem/Perfumed Garden of Exquisite Flowers and Potent Weeds ........ and as Easily Worshipped and Mastered as Such with Such.

EMC should buy Sun

For varius reasons I think EMC should consider stepping into the fray see- http://www.ecm-stuff.blogspot.com

It made sense!

It is the first time that a comment by amanfrommars made sense to me.

probably I am too drunk.

@AC

Yes the two three look set to be Power, Intel in some form and the third to ensure competition could and maybe should be Fuji SPARC. Perhaps only Power and SPARC are mainframe class RAS and Fuji ECC throughout gives excellent availability.

Pity if we end up as just two horse race Power and Intel, limits customer choice.

@amanfrommars yes, valid points JAVA is unwieldy

@Matt yes, valid great points and i'm not so much Sun fanboi as you may think!

Roll on Power and Polaris, maybe they can get it faster than Slowaris?

Cheers

Ray

End of an era

Assuming Sun is purchased by someone and ceases to exist then its passing will mark the end of an era. The era is of course the dominance of big single threaded applications and cpus. Simply adding more cores to the chip will not move the industry too much farther forward. The Niagra and Rock processor represent an attempt to adapt to the new future. If IBM buys Sun it is inconcievable to think that these chips don't contain some IP of interest and benefit to IBMs own designs. IP that will be denied to Intel and most especially the Itanium.

Java??

Yeah right, they want Sun for their customer base. Java is awful, get with the game.