HP and Novell grabbed hold of their Transitive stick this week and smacked around poor Sun Microsystems. Come July, HP's Itanium-based server customers will be able to run Solaris/SPARC applications on top of Linux via Transitive's QuickTransit software. The Itanium play adds to HP's existing deal with Transitive around moving …
That's nice, but...
Having reviewed/researched many of these "cross platform" applications to support migrating customers, Transitive as a tool is nice, but in the real world, there's no support for your typical "Off-the-shelf" application. So, this aids in migrating home grown applications, but is relatively useless for your typical Oracle, SAP, or other "common" application.
We've found that recompiling the application for the new environment or leveraging the available application (when it exists for the new environment) is the best way to go.
Also, in the best case, you can migrate an existing license to cover the application running on the new environment. In the worst case, you are forced to buy a new license. You have to compare this to the support costs that you are paying for the current environment.
In the end, I have found none of these tools to be the "Holy Grail" that some people tout them to be. Like any tool available to you, you need to properly evaluate the true feasability and ROI.
How does this tie in with HP's Solaris/x86 strategy?
I really can't see the niche for this approach, let alone that it should be mainstream. Consider:
1. If I were an ISV, I would be unlikely to support my Solaris/SPARC application running on top of Transitive.
2. If I were moving a current, supported application onto x86, my first choice would be Solaris x86 - running transitive on x86 would be very much a second best solution
3. If the app is out of support then, unless it was using non-standard API, Sun's binary compatibility guarantee means it will run on a current SPARC box under Solaris 10, in a container for first choice.
Snag for HP (and IBM) is that they hoped Sun would die, and they could hoover up the customer base. This toolset is clearly geared towards being the "post Sun" vendor of choice. Instead, Sun is going through an impressive rollout with Niagara and APL machines (and Rock now in the lab), and with disruptive innovations in Solaris 10 - and Gartner (one of the most lagging indicators of the lot) has them as "positive", with Solaris as "strong positive".
HP may get some customer's to buy into the Transitive approach, but their real problem is that they've placed big bets on GNU/Linux on x86 being a killer platform. Sun's renaissance, and especially Solaris/x86, has meant that this platform is no longer an automatic choice, and has to stand against the choices on its merits.