IBM has done a solid job of building its BladeCenter infrastructure and partner ecosystem. By “open sourcing” the architecture, IBM is positioning BladeCenter as a plug in architecture for data center infrastructure, switches, storage and so on. Something like an “Eclipse” for hardware; a plug-in architecture third parties can’t really afford to ignore from a go to market perspective. That is, why build fans, a backplane, power supplies and whatnot into componentry when the BladeCenter architecture provides these functions out of the box, and somewhat elegantly? It’s a data center in a box, for storage switches, Network switches, SSL and XML acceleration and so on. If you are an f5 Networks, a Cisco or a Mcdata then why not just provide a blade rather than a server – it gives customers a lower cost entry point for your technology, with no need to do the integration or pay someone for it. Again, think Eclipse… Competitors that claim IBM only builds tools to drive services engagements should maybe take a look at BladeCenter; its about data center plug and play. Check out the list of partners.
IBM feels BladeCenter is now mature enough it can start to sell versions designed for specific vertical industries, and the first is BladeCenter-T for telecommunications. The blade rack offers Network Equipment Building System 3 (NEBS 3) and European Telecommunications Standard Industry (ETSI) compliance.
IBM has yet to build significant momentum for POWER in telecoms. Customers, many of whom feel they took a “risc”, and got somewhat burned, by investing in SPARC, are now determined to buy only into industry standard hardware architecture; that means AMD or Intel in their minds. IBM probably doesn’t have the level of ISV support needed to build a head of steam in telco though (in the short term).
There is a simple answer. Solaris 10 blades running on AMD processors plugged into BladeCenter. This solution would potentially offer an ideal solution for telcos with significant Solaris investments. Industry standards is one thing. Porting costs are quite another. So why not stick with Solaris binaries?
Of course there is a big issue for IBM here-it claims there is no significant demand for Solaris x86, so why support it across the IBM Software Group portfolio? But from an IBM Systems and Technology Group (STG) perspective, so what? STG should build the best hardware it can to support application and operating environments, and move on. In fact STG is already doing just that in some cases; see my blog about Oracle database running on iSeries.
It may be that BladeCenter T ends up driving Solaris 10 into IBM, or should that be Solaris 10 driving BladeCenter-T into telcos? Whichever way around it is, the topology could be a winner. Lets see if STG takes up the baton and runs with it. And hey if Sun can make nice with Microsoft, why not with some parts of IBM?
IBM could theoretically be truly deeply radical, and or heretical, and use OpenSolaris as the basis for a set of BladeCenter-T offerings. It won’t happen, but customers should appreciate the potential openness on both sides. BladeCenter is well on the way to becoming a defacto industry standard. Solaris already is one.
Its time for IBM and Sun folks to have a chat – just don’t tell the COO or the GM…
No Comments