Slashdot Mirror
Red Hat, HP, Intel Join in Itanium Linux Alliance
joel_archer writes "According to this Yahoo! article, Red Hat will begin selling an Itanium version of its Advanced Server Linux in partnership with HP. This is one of partnerships currently underway between these two companies. HP is a key partner for anything Itanium-related, the company invented the design underlying Itanium before handing it off to Intel to develop and manufacture. Bolstering that effort, Red Hat and HP have signed a deal under which Advanced Server will be certified on and available with all of HP's Intel-based ProLiant servers--not just Itanium systems, but also lower-end Xeon and Pentium versions and superthin 'blade' systems."
I guess it's fine and good that Red Hat is getting in on this. But there's very little need for 64-bit desktop systems (as demonstrated by the Itanium and Alpha's consumer-market failure). 32 bits is plenty for virtually every application of a desktop computer, and will be for quite some time.
So, HP wants to officially offer Linux on all of it's Itanium, and lower servers? Itanium is going to replace much of HP's higher end server line as well if I remember my facts correctly.
This sounds very similar to IBM's linux on all IBM "backend server" offerings. You have to remember, these will be all of what used to be the offerings of both HP and Compaq when considering the market scope of this.
BTW - Oracle just matched BEA System's price/performance record for the java application server benchmark. Oracle ran with an all Linux solution on HP Proliant hardware.
HP is pulling an IBM...how interesting.
-Pete
Soccer Goal Plans
I wonder if this will affect any enthusiasm or special testing/optimisation that RedHat might have planned for the AMD side of the 64-bit fence. I can't imagine Intel not putting just a little bit of pressure on RedHat to be more forceful in their...patriotism.
I don't mean to troll, but advanced server is just now getting to 64 bit archatecture? Would someone please tell me how long *NIX has been doing this, and how far behind win-tel is?
Sigs are out of style, so I'm not going to use one...oh wait..
Pretty remarkable machine, although if you consider that software is usually almost as expensive as the hardware it's no surprise that it's a bargain. But I'll match the reliability with any of the other machines we've used in the past. Everybody agrees that it's faster than the last one (well, of course!) So yeah, I'd recommend one of these to anybody who wants to increase the overall speed and reliability of their platform without breaking the bank.
http://www.mandrakelinux.com/en/ftptmp/1024501320. d32fd091334bd166624816e3d84d319a.php#others
It looks like HP, Intel, and RedHat have been in the mix since 1999.
http://sverre.home.cern.ch/sverre/Linux_IA64_proje ct.html
Redhat is using this partnership to increase their revenues and clean up their profit margin according to this article on the Register, which coincides with this Yahoo News item.
Redhat may or may not be your favorite distro, but at least they're doing something to increase Linux marketshare, and apparently are doing it successfully.
Duris MUD - The best pkill MUD. Ever.
You 'desktop' system is probably doing quote a bit of >=80bit processing especially if you running games (the fpu),
what 64bit's gives you far more than bigger numbers, you should also get an architecture change, otherwise no-one would be bothering to develop 64bit systems, most markets that require that kind of pression are fairly well saturated.
In your example the only real problem with 16bit systems is memory addressing limit, so why didn't intel make a 16 bit processor with 32bit addressing registers? it's far easier than making a 32bit processor.
thank God the internet isn't a human right.
I have a friend at HP, I'm building two data centers in the next year, I usually use RedHat...
E| b3 D3 L337357 R007 of all baby! 8^)
My $0.02 will always be worth more than your â0.02, so
Rather that slash and burn Alpha, HP would have made all of its customers far more happy by agreeing to take EV8 to silicon. I'm sure this was in the realm of the possible.
Yes, cross-license with Intel up the wazoo and sell your employees to Intel if you like, but deliver to your customers what they need to keep their datacenters for the next decade, and also bring a stunning and seminal SMT product to market.
While we're on the subject, unifying HP-UX and Tru64 into a "TruHP" might have scored a few notches on the cluestick. Let's face it: a lot of things about HP-UX just plain suck (especially the packaging system, as Tru64 announced it was moving to RPM). HP is just beginning to implement dynamic kernel tunables and even their whole enterprise file system is outsourced. I am totally underwhelmed. When they lose the performance edge, I will have no sentimental attachment to this kludge.
Just like IBM and Sequent, HP has knifed products that work for products that don't. May Opteron be the undoing of you all.
I have installed countless proliant servers and they are very high quality boxes.
HP has it's own custom compiler for Linux Itanium. In fact HP controls the design team for the official Itanium C compiler, of which the Linux version is a showcase example. That was HP's part of the deal with Intel, to supply the VLIW compiler technology. The point is, they don't need GCC. Although since Red Hat is the de facto owner of GCC, I wouldn't be at all surprised if some HP technology eventually shows up in GCC as a direct result of this partnership.
Trust me, you didn't want HP to stick with the NetServer line of x86 servers. How would you like to be uncrating and installing 50 or more new blower fans (like I have to...), because the ones currently cooling your LP2000r dual PIII 1ghz systems are failing on average within 3 months of deployment?
or does it seem like Red Hat is kinda moving away from the little guy, getting Linux on the desktop effort, screw big corporate culture movement?
I'm not sure that RedHat was ever there. The good thing about RedHat is that they do seem committed to Open Source i.e. they actively participate in contributing to many GPL projects and don't keep anything closed. As long as they continue to do this I don't really care who they collaborate with.
RedHat still releases the software they write under the GPL, and their software is still widely available for "free." RedHat has put their money where there mouth is and is making good on their claim to charge for support and not for software. RedHat has seeded the business community with high quality Free Software, and is now reaping the benefits of their work as business start using this software and (more importantly) start paying for support.
Anyone who links Linux with some sort of lame counter-culture anti-business meme is just being soft headed. RedHat gives away software because it makes business sense to do so, plain and simple.
actually, from a business standpoint having choices is much, much better, seeing as you get the ability to tell the vendor 'meet these terms or i'll go with product Y'. With no competition, you have no leverage.
Maybe now my shares of RHAT will actually gain some value! :-) Red Hat has been making a lot of critical partnerships like this recently.
Say what you will able Red Hat vs. other Linux distributions - it is the partnerships and support of other Enterprise-sized vendors that is going to make or break Linux. Being that Red Hat is smart enough to make these partnerships, my money is on them to be "the" premier Linux vendor for the corporate market.
Take care,
Brian
--
100% Linux Web Hosting Solutions
--
Funny to see how SuSE is not part of that alliance. They were the first to ship an Itanium distribution, back in June 2001!
Have a look here if you don't believe me - this means you have to fork out $495 (yes, you read that right) for Red Hat Linux on an HP Itanium box compared to nothing extra for XP, HP-UX or indeed other Linuxes (Mandrake, Debian and SuSE all seem to have ISOs for Itanium available).
Surely HP must now resume shipping Red Hat Linux with their Itanium boxes [they did used to ship RH with the boxes until quite recently] ? Or is $495 considered peanuts compared to the cost of the boxes ?
Same with that guy who said there was only a market for 10 computers in the entire world. At that point in time, it was fairly accurate.
-
Sofware that currently uses 32 bit values to compute hard drive space is probably broken right now as the maximum value a unsigned 32 bit nubmer can represent is about aproximately 4 gigabytes. Have you ever written code to convert between an int64 and long? It's not complicated but boy can it ever be annoying.
time_t this happy 32 bit little time value is ubiquiteous accross all platforms and is also broken after 2038. Switching to a 64 bit version makes the problem go away.
32 bits starts to get a lot smaller when you're dealing with signed values. ~2billion isn't that big a number for a lot of computations.
In short, switching to 64bit will solve a lot of little niggly programming problems for free.
Operating systems with standardized ways of writing software will probably make the transition fairly seemlessly (UNIX: already available). Operating systems with rampant use of hard coded 32bit values (DWORD) for handling pointers and system resources will have a more painful transition (Windows: delayed).
Do not spread "09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0" over the internet, thank you.
HP bailed on IA-64 because HP couldn't run its development operation efficiently enough to meet the tight controls it now must use to survive.
HewPaq is no longer a frontline R&D organization, it's a computer kitbuilder.
--Blair
"Not that there's anything wrong with that."
UNIX: about a decade across their whole market.
Actually, it's a lot less than a decade for most UNIX vendors.
DEC had 64-bits first; 1992/1993 I believe, with SGI not too long afterwards. So the two guys with lowest marketshare were pretty fast out of the blocks. But where were things a few years ago? By late 1998, all the RISC vendors had at least one 64-bit piece of hardware, with half of Sun and HP's product lines moved over, IBM just starting, and SGI shipping all 64-bit hardware. But various players hadn't finished all the OS-level stuff to support that. (Source for all that here.) The transition to 64-bits wasn't done for UNIX players even 3.5 years ago, so "across their whole market" is really way too strong a statement. Wintel ran on 64-bit Alpha support long ago, but actual 64-bit APIs were still in development back in that timeframe; I haven't seen how far along they are now.
At one point in my career, I analyzed 64 bit marketing for several projects. Basically, saying "we're 64-bit, they aren't" was never a very compelling argument to begin with. Sure, in a few cases (very large databases, but not very very large databases) it made a difference, but at the end of the day, it didn't win any hardware players a lot of business.
Saying "64-bit is better" is easy, showing that 64-bit is worth paying more money is typically hard.
You're right that 64-bit Intel will likely win over 64-bit RISC long-term is right. But Intel is having huge problems executing on 64-bit Intel stuff. Itanium was a loser. We'll see how competitive McKinley is.
Right now, and I suspect for some time to come, Sun and SGI will continue to sell better hardware primarily based on "more reliable", "more scalable" kinds of features within the hardware (as usual, features requiring OS support), not leaning too heavily on the 64-bit argument.
--LP
Actually the software industry has become more and more treacherous in the last couple of years, especially for startups. For years the basic plan has been to develop a technology and hope you get bought out by Microsoft or one of Microsoft's competitors instead of simply destroyed.
RedHat has made a business of undercutting it's commercial software competitors by creating a atmosphere for collaborative software development and by packaging and supporting the results. This unorthodox method of competing has scored them some impressive wins, and it is gaining them both market share and revenue.
Contrast this with what would have happened had they decided to try and compete with a commercial product. Chances are they would end up like BeOS.
I thought Oracle's licensing doesn't allow you to publish the results of such tests... well, I'm sure they don't mind if Oracle comes out the winner...
Well, there are a number of factors that are misleading here. Unfortunately a lot of people don't get the real picture. Here are some useful things to know about Debian and HP.
1. Debian's current "Project Leader" is now Bdale Garbee, who works for HP, making sure that Debian (and the Linux Kernel) runs quite happily on the IA64 platform. Most of Bdale's changes have gone back upstream to developers, and have since filtered down to RedHat. I would not be surprised if RedHat actively watches Bdale's code changes to speed this process up. A number of other HP employees work on Debian, and in some cases, were hired by HP after already starting work on Debian.
2. Woody is the only version of Debian that will work properly on IA64, and for the IA64 architechure it uses a 2.4 series kernel. If you haven't used Woody, or used the installer, don't even bother to comment.
3. HP provide a "base system" installed on their machines. They also provide a software loader, that allows you to "choose" what version of Linux is installed. This then copies a pre-installed copy of the chosen OS over the existing system. At least on IA64, the distributions that were going to be bundled were: Debian, RedHat, Suse, Mandrake. The reason I say were, is that I last heard about this directly from Bdale at LCA 2002 (Brisbane, Australia) and there is this whole United Linux thing that may have changed what will happen with Suse somewhat.
4. HP has put it's money where it's mouth is concerning Debian. A number of HP IA64 and HPPA machines have been GIVEN to the Debian community to use as development systems and package builders. They also paid the legal bill to find out just where the legal stance is on having crypto software directly in the main archive.
5. If you want to see JUST how many packages work under Debian (particularly in reference to the IA64 architecture), you should probably look at the Debian Builder Statistics page. Approximately 94% of all of packages in Woody build on IA64. There is over 8000 packages in Debian Woody.
6. Interestingly in reference to your "support" comment, perhaps you may not know that Debian's package update for the recent Apache "Chunked Transfer Encoding" vunerability was released by the security team recently. The update message I recieved from the Apache announce list and the Debian security mailing list (both announcing fixes for the bug) were within 90 minutes of each other, and in both cases files were on the mirrors at the time the notification was released. I personally think this is pretty efficient.
I use Debian in my workplace. It works very well for us. We used to use RedHat and it caused us lots of problems, particularly with upgrading and maintenance, but also when it came to configuration. This is my personal opinion, but I believe my reasons of choice are valid. YMMV.