This morning we sent an open letter to HP’s CEO, Mark Hurd. You can find the text of it here.
I’m not necessarily the biggest fan of open letters – but every once in a while, they help explain our position, or motivate customers to engage in driving the industry. And customer involvement has led to some remarkable innovation (my favorite example being here).
With full disclosure, we’ve been trying for a while to engage HP in a dialog around converging our Unix efforts – their Unix, called HP-UX, was the second most popular Unix behind our offering, Solaris. But HP’s customers are increasingly moving away from HP-UX – not because they don’t love it (many do, and there are a lot of folks that have a deep respect for it at Sun, as well) – but because HP has limited where customers can use it. As the exclusive OS for HP’s Precision Architecture, HP has built up an installed base of more than $100 billion – for a reason. It was a great platform.
But then HP decided to end of life PA-RISC – and in so doing, left their user community with a very tough choice: if you want to preserve your investment in HP-UX, you have to rearchitect your entire datacenter to adopt Intel’s troubled Itanium project. But if you want to enjoy HP’s high volume Proliant line of x86/x64 computers, you can’t run HP-UX – unlike Sun, HP elected not to invest in supporting their Unix, HP-UX, on their own x86/x64 servers.
So we’d like to offer HP, and the HP user community, a third option: to converge Solaris 10 with HP-UX, running on HP’s very own Proliant product line. We’ve spoken to HP about it, thought we saw a glimmer of interest, and now we want to get their customers and partners involved.
To build up to this point, we’ve been doing our best to faithfully support Solaris 10 on HP’s entire Proliant family of servers. They’re all qualified – try the download, here.
And we’re going to continue extending the olive branch to the HP-UX user community, with tools, technologies and partner support. But we’d like HP to contribute to the effort, on behalf of and in concert with their customers and developers.
So now it’s up to you – the HP user and developer community. Talk to your HP sales rep or management contact if you’re interested in seeing such a project flourish. You have all the power in this dialog. And Sun would be very supportive – I commit that. You shouldn’t have to abandon HP-UX, and we’ll gladly work to ensure it.
At bottom, a converged Solaris/HP-UX roadmap would offer our joint customers choice, innovation, and a broad spectrum of qualified hardware – while preserving your investment in skills, your confidence in the underlying intellectual property and technology… and most importantly, given that Solaris is open source, your choices going forward. That’s all upside, as far as we’re concerned.
Now we just need to get HP on board…
What I Couldn't Say... 
Since HP and Intel and other itanium creators have ponied up another $10 billion dollars for itanium, you should offer Sun’s design services to them, since obviously Sun Labs are better at new CPU design since, they created the UltraT1 that puts them to shame on some work loads, and did it all for what a few hundred million max? Offer to do it for just 5 billion dollars, and they can save 1/2 them of the fresh input of cash they decided donate to there cause.
10 billion more dollars for itanium
The open letter is already generating the publicity you might expect. As usual The Register puts thier own spin on it predicting that this might mean a port of Solaris to Itanium. But then as The New Yorker pointed out this morning, in a review of a book by Louis Menand, predictions are very hard to get right.
What precisely do you mean by “converge?” One assumes that your proposal would have the ‘converged’ OS drawing a lot more of its genetic material from Solaris than from HP/UX. Are we talking about binary compatibility for HP/UX applications? I don’t think anything short of that accomplishes your stated purpose (providing a lifeline to loyal HP/UX users). If these users must be required to switch to binary-incompatible OS, well, they can already do that today – with Linux.
Aside from binary compatibility, much of the pain of switching OSs revolves around retraining humans, and retooling scripts and other infrastructure pieces (e.g. customized OS-specific monitoring) to accomodate the new OS. Again, Sun needs to show how a converged Solaris/HP-UX would make this less painful than a wholesale switch to Linux.
It seems the onus would be on Sun here to clearly illustrate how switching from HP/UX to Solaris on HP hardware would be preferrable to switching from HP/UX to Linux on HP hardware. I know Sun feels that Solaris is a better enterprise OS than Linux, and I’m personally inclined to agree, but we seem to increasingly be a minority, and this is the case that Sun needs to make in order to redeem that open letter as anything more than an inflammatory marketing maneuver.
Why don’t you converge on Linux instead of starting competing open source projects?
To Penguin,
Why can’t there be competing open source systems ?
The smarter (and cheaper ) ones will have more mindshare. Let innovation take its path.
Note: RedHax linux is proprietary.
HP lied to the Tru64 UNIX customers when they claimed they would port various Tru64 technologies such as TruCluster and AdvFS to HP-UX.
Why should I trust the word of HP?
Fortunately Linux has freed us from the stranglehold of proprietary UNIX.
hi,
how u doin?
I am a student studying Business IT Systems (MSc) at the University of Strathclyde in Scotland.
For my masters dissertation I am investigating the relationship between commercial companies and the open source community. The project examines the issues, motivations and critical success factors in this collaborative relationship.
I was wondering if you could take part in my survey.
The survey can be found at … http://FreeOnlineSurveys.com/rendersurvey.asp?sid=v4qpja3izsf7w9f170254
thnx,
imran
Why can’t we all just be friends?
Prince – You are right in that the more open alternatives the better – it leads to innovation. But as a matter of fact, you are DEAD wrong on the RedHat is proprietary statement – Redhat is only as proprietary as Solaris is today – You can get the source for all Redhat software and you can even download a trial which doesn’t expire and is legal to use commercially – only thing is Redhat will not support it should you run into issues. This is same as Solaris/OpenSolaris.