Hewlett-Packard is facing a difficult time in the server market. Indeed, IBM is gaining momentum amid HP’s missteps with Itanium, the latest Gartner research suggests. Now, HP is trying to save face with a server initiative code-named Odyssey. But here’s the big twist: Odyssey’s biggest potential winner is Red Hat. Here’s why.
At the high-end of the server market, HP recently revealed that its Itanium server sales are plummeting. HP blames the weak Itanium server sales on Oracle, which has abandoned Itanium-based software development. Unfortunately, HP fails to concede that Red Hat and Microsoft had previously abandoned Itanium, so HP should have seen the writing on the wall for the high-end processor. Meanwhile, Dell is striving to attack HP in the more traditional x86 blade server market.
HP’s Latest Journey
Here’s where things get interesting. HP is now evangelizing project Odyssey, which will strive to unify the Unix and x86 server markets. (Side note: The VAR Guy heard the same story nearly 20 years ago… it was called Novell UnixWare and it failed. Oh, and then their was Solaris for x86, and that failed too.)
HP insists it remains committed to Itanium, but read between the lines and Odyssey is a lifeline for Itanium customers who want to migrate to Windows- and Linux-based servers over the long haul.
And that’s where Red Hat enters the picture.
Predictable Strategy
Bowing to market trends, HP says its Odyssey strategy will allow customers to run Windows Server and Red Hat Enterprise Linux (RHEL) on HP’s highest-end servers. Um, that’s a 10-year-old trend. For current and former HP-UX customers, the most logical migration strategy is from HP-UX to Red Hat rather than Windows, since Unix and Linux are close cousins.
Whether Odyssey succeeds in a big way or merely limps along, the company that stands to gain the most is Red Hat — which built its business on Unix-to-Linux migrations.
Read More About This Topic
Share This Post
Tags: HP Itanium | HP Project Odyssey | HP Servers | HP vs IBM | HP vs Oracle | Red Hat Enterprise Linux | RHEL
Interact: Add a Comment | Trackback Link | Permalink
Subscribe: RSS Feed
Don't miss Charlene O'Hanlon's weekly columns...
Hm. This is a bit odd sounding. Only odder thing would be if they had said it will run HP Linux, a special Itanium-optimized, HP-UX super-kernel layer, and will revolutionize computing. Or that it would come with an 8080 chip emulator to run CP/M applications also.
Seems like they have customers who are telling them “all the major vendors have given up on Itanium, and x86 with Linux systems are running almost all enterprise mission critical workloads, as well as all new Internet workloads and applications, so we are planning to migrate off our HP-UX box as soon as feasible”. Which is logical, and which is the overall market trend for the last 10 years from UNix moving to Linux. This part of HP must now feel like Digital, Prime, Data General and Wang did at one point in computing history. You are correct that Red Hat, and people selling applications that run on Red Hat, are the continuing winners in this game. From what I can see Red Hat has the only operating system offering that is both widely proven and accepted, and yet is still hip and modern – a rare thing these days.
Bill,
The VAR Guy thanks you for your continued readership. Always good to have a dialog with you. Agreed: Odyssey sounds a bit odd to The VAR Guy. Sounds like HP wants to take all the best capabilities of HP-UX, NonStop and other technologies, and somehow work them into x86 servers running Windows or Linux.
Give HP credit: It needed to give Itanium customers an alternative path forward. Odyssey may fulfill that need — though HP should have seen this day coming far sooner than November 2011…
-TVG
I don’t see how Odyssey does anything for current HP-UX, VMS, Nonstop users that are dead-ended. Basically all HP has said is that they are going to build high RAS x86 boxes. It is still a completely different chip set, different compilers/binaries, different OS, probably different systems management tools. All of the legacy HP OSs will remain on Itanium until the end of the line…. This is no different than the path an HP-UX customer would have in migrating from HP-UX to Linux on IBM System x or Dell. The enhancements HP is bringing to Linux from HP-UX are going to be on the market “within two years.” That isn’t going to fit the timeframe of most HP-UX users who need to migrate yesterday.
On the Red Hat point, I wonder how all of the Oracle users on HP-UX, who are really the group in a tough spot, will feel about migrating to Red Hat vs. Oracle Linux given the on-going battle between RHEL and OEL. How long will RHEL be supported for Oracle applications?
Sam,
The VAR Guy appreciates all the points you raise. Our resident blogger believes Oracle will certify applications for Red Hat Enterprise Linux for years/decades to come because RHEL has a critical mass of customer adoption. Remember: Oracle makes big money selling databases and apps on both Windows Server and RHEL. No reason to abandon those revenue streams unless Oracle builds an even better, more successful Linux platform. So far, Oracle’s Linux alternative is a niche offering, The VAR Guy believes…
-TVG
“Remember: Oracle makes big money selling databases and apps on both Windows Server and RHEL.”
Yes, but that was true of HP-UX too. If Oracle would have left the Itanium multiplier at 1.0 instead of cancelling development, they would be printing money on that platform with their DB licensing. It is possible that Oracle will want to cool things down after the HP-UX war, but that is not really their style. They could care less about the Windows platform. Those are small nickel and dime installs. Linux is becoming their enterprise standard. They are incentivizing customers to use OEL and OVM by wrapping it into ULAs and punishing users that make the wrong decision (RHEL and VMW)with support and licensing complexity.
HP should acquire EnterpriseDB. Build these Linux servers with high RAS, bring the HP-UX IP to Linux and then start wrapping PostgreDB with the Oracle DB migration tools into the support costs on these boxes. They would need to beef up Postgre’s concurrency controls, data types, scalability and other stuff to compete with Oracle, but providing a completely open, license free high-end server with an OS that is pretty close to Unix and a DB that is pretty close to Oracle would be interesting. I am not holding my breath though.
Regarding: “Side note: The VAR Guy heard the same story nearly 20 years ago… it was called Novell UnixWare and it failed.”
Yes, and Ray Noorda told Novell they should go Linux instead back in 1993-1994 as well. Novell didn’t realize Noorda was correct for another eight (8) years.
Hi Sam,
Isn’t OEL a RHEL-clone? OK, they have their own kernel, optimized for their DB. But the rest of the OS is mostly recompiled RH source code, afaik. So Oracle relies on RH and cannot kill them (they could buy them…OK). And for the punishment: I wouldn’t install a DB-Server in a VM. I’d rather buy a realy small box and install it on the iron.
Even Oracle’s own kernel is still based on Red Hat developments. E.g., where does one think the 2.6.32 kernel came from?
As far as Oracle buying Red Hat, I think IBM might have some say about that. That’s 100%, continual investor speculation with little merit.