HP preps Tukwila servers for April 27
Bearing witness for Itanium
We've all been wondering where Hewlett-Packard's high-end servers using Intel's latest Itanium and Xeon processors have been hiding. Well, it looks like the Itanium boxes will be announced on April 27.
The company is being tight-lipped about exactly what it has planned for the April 27 event, but it appears to involve balancing an enterprise-class server atop the full belly of a dinosaur that has keeled over dead:
OK, perhaps that is actually an Integrity server riding the back of a dinosaur, metaphorically a mainframe, or maybe that image is supposed to be a mountain peak, in reference to the Superdome brand used by HP's high-end 64-socket PA-RISC servers and their Itanium kickers.
Anyway, the HP invite for the April 27 extravaganza says that Intel, Itanium, and Itanium Inside are Intel's trademarks, without ever mentioning them, so we know this is the long-awaited (well, for HP-UX customers anyway) announcement of Integrity machines using Intel's quad-core Tukwila Itanium 9300 processors. The Tukwilas were launched in February at the International Solid State Circuits Conference, with only Intel and HP on hand to talk, and at the time HP said that it would get Tukwila machines out within 90 days.
The April 27 launch of Tukwila-based Integrity machines coincides with the HP Technology@Work 2010 conference, which is being held from April 26 through 29 in Frankfurt, Germany. So is another related event called the HP Enterprise Technology Summit.
HP is not letting the cat out of the bag specifically in the Technology@Work session schedule, but there is a NDA session (BB-60) that talks about the "newest 8 socket scale-up workhorse," which has to be the ProLiant DL980 El Reg already told you a little about two weeks ago. The DL980 will use Intel's eight-core Nehalem-EX Xeon 7500s, of course, which employ the Boxboro chipset from Intel and the buffered memory card and I/O architecture that was designed to be compatible with the Tukwila Itaniums.
Session BB-17 at the conference in Germany covers the new Integrity servers, for which HP says "the clock has been reset for a new decade of mission-critical computing." This session, HP continues, "will cover the value delivered by New Integrity Servers including product overviews, business benefits and total cost of ownership comparisons." Yet another session, TB-04, will "cover New Integrity HP Superdome Server in detail, including its completely modernised mission-critical design, massive scaling capabilities, and world-class availability features."
Martin Fink, who runs HP's Business Critical Systems business unit, will give a 45-minute keynote on mission-critical computing and explain why HP has invested in an "entirely new range of Itanium-based Integrity Servers designed to combine the benefits of latest technology developments and HP's mission-critical operating environment." Fink will also explain how the new Tukwila iron "will make your mission-critical data centre future-proof."
I don't mean to be mean, but that is truly funny. Unless future-proof means something other than what I think it means, and unless you don't know the 15-year history of Itanium and its disappointments. It would be hard to find a platform more future-proof outside of a mainframe. As long as Intel is making money, it can indulge in HP's Itanium habit, which is good for HP-UX, OpenVMS, and NonStop customers that don't want to port their code to another platform.
The real interesting bit will be what HP does for chipsets in the high-end Superdomes. Any entry or midrange machines can be based on Boxboro chipsets and basically globally replace Xeon 7500 sockets in a ProLiant DL980 with Itanium 9300 sockets, and slap an Integrity Whatever name on it and a much higher price tag. With the Superdome high-end machines, HP has created its own chipsets, with the most current ones being the Arches sx2000. HP has said absolutely nada about the kicker to Arches and really doesn't seem inclined in the past several years to talk much about Itanium - unlike in years gone by. ®
Where is he?
Conspicous in it's abscence, yep, thats the lack of Itanium support. RHEL + Microsh!te dropped Itanium so it must be considered a real specialists market, perhaps this announcement is nothing more than the kick of a dying beast?
Matt, as our local HP rep how long till HP engineering put their VMS / HPUX customers through another code port and another load of ISV's just don't bother to recompile? How small is the ISV App list on HPUX these days comparted to Linux, AIX or Solaris?
My money says VMS will code port to x64 within a year and thats assuming HP doesn't have engineering on it already. Dropping HPUX for Linux is easy but people fixate on VMS so it probably has more life.
one last 65nm hooray
Intel can't wait to kill this line as it has become the equivalent of the Cadillac Cimarron (a proud brand with a terrible mistake blighting its good name). Intels knows that the IA64 was a terrible idea with even worse execution that due to them not listening to their customers almost cost them dearly against AMD (thank God for chipzilla Hector Ruiz rode to their rescue and ran AMD into the ground) The chip has only made it this far because Intel had to meet its contract agreements with HP for all the tasty DEC IP they lifted. Just seeing Intel putting all the RAS stuff from IA64 in the Xeons should tell everyone but the locked in HPtards (hello Matt B. you there) that the end is nigh. By the way IBM Powertards you are next.
RE: @MB? FUD Who?
"....Is this FUD or what?..." Erm, yes, that was the whole point! I was citing examples of carefully prepared FUD that I have seen in the industry regarding IBM kit, to show how my own amateur examples were nothing by comparison. D'uh! Did you even read my post before you went to Auto Squeal Setting 10?
".....Being an IBMer...." Ah, another unbiased and equitable viewpoint then. Please do shower us with your no-doubt extensive experience of anything non-IBM. Yeah, I'm not holding my breath on that one!
"....I would love to see proofs of this BS!...." I suggest you call hp then. It is with an evil smile that I hope they subject you to the full death-by-PowerPoint experience they give us customers, as it would seem fitting to have a salesman suffer some of their own medicine!
"....You don't like FUD , but you spread FUD when it is convenient!..." No, I don't like FUD, so when you IBMers (and Sunshiners) start up with it I feel it is only fair to throw some right back. Moral of the story - trolls in glass houses shouldn't throw stones if they don't like the sound of breaking glass.
"....HP is years behind POWER...." A truly eloquent, concise and balanced argument without any hint of ranting. Nope, no ranting at all. Nada! If only all IBM salespitches were that short! Mind you, and please don't take this as a criticism, but it is a bit lacking in technical information or any sort of reasoning. Well, actually it's short of anything other than unproven supposition masquerading as fact. But, as a first effort, it does at least leave plenty of room for development into maybe a whole sentence. I suppose it was far too much to expect anything like empirical facts or industry analysis to back up your statements but I'm sure that one day you'll be able to supply some when IBM release some more brochures for you to refer to. Who knows, one day you may even communicate an orignal thought!
"....IVM is a joke. Don't know how HP calls it virtualization with so many limitations...." I suppose we have to at least pretend you know something about IVM that didn't come out of an IBM FUD guide, so please elaborate by explaining what you mean by "limitations", and how it fails to compare to IBM's virtualisation offerings? Please do bear in mind I've already debunked the standard IBM FUD as offered by Ms Park (http://forums.channelregister.co.uk/forum/1/2010/02/09/intel_tukwila_feeds_speeds/), so simply repeating the same FUD again is just going to make you look even more unoriginal. Postively sheep-like, tbh!
I see the IBMers still haven't taken my suggestion and pooled resources, rather than just resorting to the same IBM FUD guides. Bit of a disappointment, really! Maybe the Sunshiners can lend them a hand?