Microsoft and Red Hat agreement and the next generation datacenter
Microsoft and Red Hat have come to an agreement allowing virtualization technology from each company to support virtual machines containing the others operating environment. My colleagues Mary-Jo Folley (see Microsoft and Red Hat sign virtualization pact) and Paula Rooney (see It’s true: Microsoft and Red Hat cooperate on virtualization) have both comment on this move. I'd like to look at it from a slightly different perspective (so what else is new?). I am viewing this event from the vantage point of organizations wanting the implement their own next generation datacenter, an important step to creating their own in-house cloud computing environment.
As I've pointed out in previous posts (see What would a next generation datacenter look like? and HP and the next generation data center), a next generation datacenter would have a unified industry standard hardware platform that supports a diverse set of workloads, each of which might run a different operating environment. In that datacenter, workloads running on Windows, UNIX, Linux, Mac OS and possibly even OS/400 and Z/OS could all be at the party.
It appears that the agreements between Microsoft and both Novell and Red Hat clear the way for Windows and Linux to be at the party. Sun would point out that their xVM project is, in effect, an invitation for Solaris to be there too. This means that hardware suppliers such as Dell, HP and IBM will be there. Apple's current licensing rules make it difficult for Mac OS to be there too. Apple doesn't like to have its Mac OS living in a virtual world.
IBM faces some challenges that IBM's vendor specific hardware architectures appear to make it difficult for either OS/400 and Z/OS based workloads to be at the party.
IBM's recent acquisition of Transitive might offer a clue at the thinking of good ol' big blue (see IBM Acquires Transitive). Transitive's technology makes it possible for native HP-UX and Solaris/SPARC applications to be run on that unified industry standard hardware platform. It is conceivable that similar technology could be developed (or perhaps has already been developed) that would give a ticket to IBM's OS/400 and Z/OS.
Do you think it likely that IBM will print that ticket making it likely that it would sharply reduce sales of its mainframe and midrange hardware platforms? Would the revenues from increased sales of its System X make up for declining sales of its System Z, System I and System P?