Last week, I received a notice from IBM's PR company that IBM was going to acquire Transitive, that clever supplier of operating system and hardware virtualization technology. It's clearly a move to strengthen IBM's claim on being the premier supplier of systems for the next generation datacenter.
A while back, I posted some ruminations on the requirements for next generation datacenter (NGDC) and how they related to systems design (see What would a next generation datacenter look like?, HP and the next generation data center and IBM announces the Z10: Is the mainframe still relevant?).
One of the major requirements of the NGDC is software than can allow a diverse set of workloads (meaning designed for different operating systems and processor architectures) to run on a unified hardware architecture.
At the time those posts were published, I pointed out that HA and IBM were bringing out systems that were clearly designed to be the hardware platform for the NGDC. Each of these competitors announced the ability to bring workloads built on some combination of Windows/X86, Linux, Solars/SPARC, Solaris/Intel, HP-UX/Precision, and AIX/Power onto their systems and have them execute side by side. As you might guess from the acquistion announcement, the magic behind this capability often was supplied by Transitive.
Transitive has been a victim of, oops I meant a target of posts here on Virtually Speaking in the past (see Transitive QuickTransit - Hardware Virtualization that Takes the Pain Out of Migrations
, Transitive’s magic - Solaris/SPARC Apps and Hyper-V
). Their technology reminds me of some products created by Digital Equipment Corporaiton, now part of HP, that allowed VAX/VMS software to be re-hosted on Alpha/VMS systems and VAX/ULTRIX software to be re-hosted on Alpha/OSF 1 systems without changing a line of code of those applications. Transitive appeared to be the only company pursuing this line of industrial sorcery.
It's clear that IBM needs to be a major player in the market for systems for the NGDC as it has been for both past and present datacenters. The company saw that getting mainframe, midrange and industry standard workloads to co-exist happily was becoming an absolute requirement. It clearly feared that without such technology it could be left behind as organizations moved from diverse hardware platforms running a diverse workload to a unified hardware platform supporting that very same workload.
The only way for IBM to gracefully get to that position from its multi-hardware/multi-operating system position today was building or buying something like QuickTransit from Transitive.
This acquisition will make it easier for IBM to capture the leading position in the next generation datacenter. This acquisition is not without risk, however. IBM has acquired companies in the past only to watch the key people who made the magic happen leave because they didn't like the IBM culture, the requirement to move to a new location, or some other reason (anyone remember IBM's acquisition of the telecommunications supplier ROLM?).
Another challenge is maintaining Transitive's reputation of a neutral supplier of technology. If the company becomes to IBM-focused, its tecnology will not be seen as the bridge technology upon which the next generation datacenter ought to be built.
We'll all have to see how this works out. At first glance, however, IBM's move appears smart.