Is it time to ditch HPUX?

[unixdaddy]

I'd like to get some comments on the attached link. It is obviously biased, but it does make some thought provoking comments. What future does to HPUX if people don't buy itanium in the future?

http://www.sun.com/executives/realitycheck/reality-091004.html

[Kent Ostby]

Darren --

There are several critical overstatements in the article.

For instance:

------
"IDGNS: Would you think of porting HP-UX, for example, to the x86 platform, now that there are these 64-bit systems?

Livermore: No, we don't think that we need to do that."

This statement from HP's number two executive should send shivers down the spines of HP-UX customers, because it categorically states that HP is not prepared to make HP-UX available to one of its largest platform bases
------

Actually, she didn't say that at all. What she said was that "we don't think that we need to do that"... she didn't say "We will NEVER do that" which would have been catagorical.

Basically, what you are reading is Red Meat posted on Sun's website for their employees.

I also have trouble taking seriously an article which contains the words "highly regarded Solaris" but then thats my HP bias showing through.

Best regards,

Kent M. Ostby

[John Poff]

Hi,

A wise businessman once told me that if your enemies are talking bad about you, it is because you are hurting them. ;)

JP

[Patrick Wallek]

I would pretty much ignore this as Sun B.S.

HP-UX will be around for a long while yet, I think. Granted PA-RISC will be going away, but Itanium will still be running. My guess for the future would be that Itanium **MAY** replace the x86 chip lines, at least for servers, eventually.

I wouldn't worry too much about some of the "stuff" that Sun spouts.

[harry d brown jr]

Of course www.sun.com could soon be renamed www.kodak.com.

http://www.10nbc.com/news.asp?template=item&story_id=12570

live free or die
harry

[Rick Garland]

It appears that SUN is making a survival statement.

[Victor Fridyev]

Hi,

The source is not neutral. Ask HP about SUN, and the answer will be much more friendly 8 ))).
I saw an HP presentation re HP HW/SW roadmap. The situation is not very clear. AFAI understand, HP intends to work some years with PA-RISC machines and HPUX11i until an ITANIUM version will be more acceptable. The transition should be smooth:
PA-RISC with HPUX11 -> PA-RISC with HPUX11i -> ITANIUM with HPUX11i. (HP`s opinion). Let's wait and see.

[Peter Leddy_1]

I wouldn't really be worried at all, it would seem just a tad bit ridiculous of HP to do away with HP-UX when they have spent millions in developing it and in my view will be the most serious contender to be the #1 Unix in a couple of years especially when the Tru64 enhancements are added.

[Rita C Workman]

What's the old saying believe nothing that you hear, and only half of what you see....

Seems that most folks (and business) just like to blow alot of smoke.

Wait till you can feel the heat, then you know which way to go.
...and I don't see any flames from my view.

Just my 2cents,
Rita

[Pete Randall]

Darren,

This sounds remarkably similar to a discussion from last week:

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=692569


Pete

[Eric Antunes]

HP-UX is the market leader, the rest is "paysage" like the French says!

Regards,

Eric Antunes

[Geoff Wild]

Ah yes - one of the Sun admins pointed that to me last week....

Just another example IMHO of a dying comapny's way to try to reclaim lost market share...remember SCO?

Attached is an 11i roadmap fro a recent presentation.

Rgds...Geoff

[RAC_1]

I am worried a bit about the job. With just HP-UX with me, I am open to accept the some facts. While the article is biased, there is no doubt that Itanium has not succeded to the extent it was forecast.

I always wondered why HP-UX, was not available for Intels and what the hell was reason to end the PA-RIS product line.

I am waiting for the one, that would run on PA-RISC, Itanium, PA-RISC and Intel, AMD...

That is my view. Hope I see a day when this happens.

Anil

[Mark Greene_1]

Darren,

Not only do I not agree with the conclusions of the article, I am rather surprised they published it, given that HP can, if they desire, take legal action against Sun over it:

http://pubs.acs.org/hotartcl/chemtech/97/jan/bad.html

[Steven E. Protter]

No.

HP-UX is a world class, enterprise capable OS. I work with Linux and HP-UX, HP-UX and the tools that HP ports to it are much higher quality out the door than Linux.

My recent series of issues with Samba which were all traced back to poor Q&A at Redhat are good examples to look at.

PA-RISC is going to be sold and supported for years.

By the time HP-UX v3 comes out in 2006(not speaking for HP, parroting what I was told) you'll have a unified code base and it won't matter what hardware is running the OS.

SEP

[Zinky]

The name of the game is "commoditization" guys. If you do not offer something that is not "commoditized" - then you're headed the fate of Dinosaurs. Just as IT had become commoditized (read: Does IT Matter by Nicholas Carr) -- so must we as admins need to prepare in a commoditized world. One must be skillful not only in one flavour of UNIX but all flavours as well as other OSes.. Already "groupings" of admins are slowly merged with no more designations of HP-UX Admin, Solaris Admin, AIX Admin or Linux admin.. hey big corporates will save on Group/Director leads on these erstwhile useless groupings!

Now is it realy time to ditch HP-UX and is it really doomed? I think this will largely depend on HP -- if they're able to sustain R&D dollars to maintain it for the Itanic err Itanium platforms - then it should survive. Else, they should commoditize the OS -- make it work on non-Integrity Itanium systems and better yet port it to other X86_64 platforms..

My few cents.

[David Child_1]

Okay, where do I start;

Statement: "Third, it allows other operating systems, such as Linux and Windows Server to establish a foothold on the Itanium platform well before the new HP-UX will be available."

Response: The key words in the statement are "new HP-UX" (i.e. v3). HP-UX is already running on Itanium and will continue to. So for author to imply HP-UX will be offered on Itanium when they get around to it is just spin.

Statement: "The largest installed base HP has is its x86..."

Response: Yes, but this is primarily thier Windows business. What are they trying to say here? Is HP suppose to just pop to and port HP-UX to x86 just because 55% of HPs business is based on Windows? It doesn't make any sense. As a side note, does anyone know how much of Sun's business is x86-Solaris based?

I had others, but it would be a waste. What I want to know if Solaris, which "provides customers with flexibility and choice of hardware platform" will have decent volume management software in its future (or must we continue to make that extra Veritas purchase).

Just so you know, I am equally responsible for HP and Sun servers. Our Sun servers are definately not as stable as HP's (strictly hardware speaking).

David

[Geoff Wild · Accepted solution]

Just received this from HP, I assume it is okay to post.


Rgds...Geoff

[Sundar_7]

With consecutive falling revenues for the last double-digit number of quarters, it is about time Big Blue acquires SUN and gives all of us a break from these kind of no-sense reality checks. :-)

[dirk dierickx]

It might be time to ditch HPUX in a few years, but not for Solaris at least.
I think all the unix versions will die out, and Linux might replace them all, or at least a very big part of them that it will be the dominant unix variant.
HP won't care to much about that, they are already confortable with selling servers with Linux and support etc. (IBM and SGI are there already too), but SUN still hasn't discovered it and unless they do, i can't see a bright future for them.

[Tor-Arne Nostdal]

In our organisation we are also very eager to see how it turns out.
We primarily use our HP machines for running SAP systems and Oracle.

PA-RISC with HPUX11 -> PA-RISC with HPUX11i -> ITANIUM with HPUX11i.

We have walked this line, and have come to installing Itanium for application server purposes, but currently not as DB-servers.

Our question now is to which extent it is possible to head further also into the Linux world.

This will give us even bigger advantages when comes to scalability, portability etc.

Does anyone have comments to this?

/Tor-Arne

[Steve Lewis]

My opinion is this:
HP produce the 'Rolls Royce' of enterprise server in Itanium and PA-RISC, together with the best operating system in HP-UX.

Unfortunately our organisation like many others is run by accountants, who rarely listen to techies because they dont appreciate people who aren't like them.

So they make up their own minds based on number of cpus, memory, storage, so on the face of it, other manufacturers boxes can appear to come out cheaper, even including cross-training, especially when you run a couple of COTS packages so have little in the way of porting costs.

I am having to fight off a cost prompted move to someone else for our new servers even now. They don't appreciate that double the CPUs at a higher clock speed is not necessarily quicker. However the glue that sticks us to HP has more to do with the support levels and the operating system than the cpu. How many organisations code in assembler these days anyway?

Interestingly, the problems in porting HP-UX to any other architecture can be gleaned from perusing the PA-RISC assembler guide. The PA-RISC hardware and o/s are heavily inter-linked and therefore inter-dependant, with dedicated registers for shared memory, multiple privilege levels and other features a bog-standard RISC chip will not have.

Operating systems aside, HP and Intel positioned the Itanium 2 systems at the top end of the market and may have to re-visit this in time, given the strong price/performance of AMD Opteron and even Athlon64.

[rmueller58]

If the competition is bad mouthing it generally is an indication you are taking some of their customers.. Reality Check for Sun. I've touched most of the major Un*x vendors at one time in my career SCO, AIX, DG-UX, SunOS, (Linux, slackware, redhat) and HP/UX.

For companies driven by customer relations, HP and it's product line is the best of the heap as far as response, and available assistance. That is my take on the whole thing. ANY dealing I've had with Sun left a bad taste in my mouth. Coming from a guy in the trenchs that does have time to keep up but to manage my systems needs quick response. Screw the marketing junk, it means nothing.

[Ted Buis]

HP-UX if ported today would take years to accumulate software on the x86 platform. The proof is that Solaris doesn't have a significant ISV software base on the x86 version. By that time, Itanium will be offering twice the performance at cost parity to x86. Solaris for x86 has lost the war to LINUX. Note that Solaris on SPARC is big endian and Solaris on x86 is little endian. This makes for a hard transition between the two. Itanium is bi-endian, big endian for HP-UX and little endian for LINUX.
Much better roadmap. What will die is SPARC.

[David Child_1]

Another note on our Sun/HP "wars". In the past few months we have dropped our Sun support and HP has taken over. So now HP supports our Sun and HP boxes. This was due in part to cost, but a larger factor was the level of support. Our HP support has far outshined the Sun support.

We are also currently porting a 1TB database from a SunFire 6800 to an nPar on a SD3200 SuperDome. The DBAs were unable to crank out the performance they needed so we migrated it to the SD32000 (approx same CPU/memory) for test purposes and it ran much faster. We are in the process of making that a permanent move.

David