From: Robert Honore (robert@digi-data.com)
Date: Thu Dec 19 2002 - 09:29:31 EST
Dear Fellow Managers,
I had received several responses to my questions, for which I would like to
thank those who replied.
Basically the advice was as follows.
* It depends on your applications, whether you are still running V4.0x or V5.1x,
whether you are dependent upon third-parties for support of key applications or
technologies, or whether you have a problem being an early-adopter.
* Keep the current platform (Tru64/Alpha) until you must move.
* No move before 2005 is really necessary.
* Use the VAX-to-Alpha transition as a guide to how things will play out.
* Right now the only systems for which one should specify HP-UX/Alpha are
entirely new applications which have no need for clustering or ADVfs.
* An important caveat is application support. If the application's supplier has
committed to provide new solutions for Tru64 for a long time into the future,
then it is best to stick with Tru64/Alpha. If your application's supplier is
unwilling to commit or is unresponsive, then the best advice is to begin work on
the migration sooner rather than later.
* HP-UX/Alpha is all-new hardware and software, so expect bugs.
An important piece of advice came from Selden E. Ball Jr.
"Converting to HP-UX is going to be as big an effort as converting to any other
non-Tru64 version of UNIX. There will be far more incompatibilities than anyone
can anticipate at this point."
* Tru64 UNIX was found to be technically superior to HP-UX and some managers
(myself included) do not look forward to having to revisit that question.
* Some have already started migrating to Solaris/SPARC and Linux/i386 (or other
non-HP platforms) and they are finding that process a struggle.
* If you are happy with Tru64/Alpha then stay there until they (HP) prove they
have something better. That will take more than just a roadmap.
Finally, I include verbatim the advice from Thomas Blinn of HP, as he, as usual,
offers very salient advice.
> You know that any valid answer begins with "it all depends". Some of
> the things it depends on are what you are running for applications,
> whether you are running V4.0x or V5.1x, whether you are dependent on
> third-party vendors for support of key technologies, whether you want
> to be an early adopter or move when you can no longer keep your old
> Tru64 UNIX systems running (e.g., get hardware maintenance).
>
> Our public commitment is to continue to manufacture and market Alpha
> hardware systems that run Tru64 UNIX through at least 2006 or there
> abouts, and to continue to support Tru64 UNIX on Alpha systems into
> 2011 or there abouts, provided there is continued customer demand.
>
> The current generation Alpha systems (the new GS1280 EV7 based box
> and the related ES80 and ES47 systems) are quite competitive in the
> market from a performance and price perspective. We have a lot of
> early orders, and expect to have a very profitable continuing Alpha
> business from these systems. There will be a product refresh with
> the EV79 processors during the life of these systems. I personally
> doubt that the migration of TruCluster technology into HP-UX will
> go as smoothly as everyone hopes; in reality, we are still finding
> and fixing problems in the V5.1B implementation, and we expect to
> continue to improve the quality and stability of this key software
> on the Tru64 UNIX / Alpha platforms for some time to come. But we
> are not going to port Tru64 UNIX / TruCluster in its current form
> to Itanium.
>
> People are expecting the next generation Itanium systems (the ones
> about 6 months or so away) to begin to be viable platforms for at
> least some commercial applications. But what you can get right now
> simply doesn't match the performance and scalability of either the
> Alpha based systems or the PA-RISC based systems. So it will be a
> while before there is a compelling reason to move, unless you want
> to be an early adopter or you are dependent on some third-party
> applications that are available on both Tru64 UNIX / Alpha and on
> HP-UX / PA-RISC today (and will be available early on Itanium) and
> your software vendor is pressuring you to move (and prepared to
> help you with the data migration).
>
> My call: Tru64 UNIX / Alpha and TruCluster aren't broken, unless
> you have a compelling reason to move sooner rather than later you
> should have several years to plan and decide (and keep an eye on
> Linux on Itanium unless you really need the TruCluster technology,
> it is actually more compatible in some ways with Tru64 UNIX / Alpha
> than is HP-UX / PA-RISC today or HP-UX / Itanium long term).
>
> Tom
And my original questions:
>
> Dear Fellow Managers,
>
> I have to produce a report for my boss regarding the issues or gotchas with the
> Alpha to Itanium transition. I thought I would solicit your opinions before I
> proceed. My questions are as follows.
>
> How do we decide the appropriate or best time to begin the transition?
>
> What issues should motivate such a change?
>
> What potential pitfalls do you see, if any? This might be a pre-mature
> question
> given that the merging of features of HP-UX and Tru64 UNIX have only just
> begun.
>
> In your honest opinion, how long do I have with Tru64/Alpha again? They
> really have served well.
>
Yours sincerely,
Robert Honore.
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:49:02 EDT