HP OpsWare, a logical, pre-emptive move

As many know HP announced it’s intention to buy OpsWare for $1.45B on Monday (7/23/2007). Some seem think that they are just buying into a negative cash flow business. My gut, tells me that OpsWare has had problems because deeper management software integration (OpenView) and HW integration would help them to create multiplicative value in operational efficiency; easing the “we’d buy it if it were just a little more integrated” denial.
So let me just play devils advocate… HP has been investing in virtualization/grid technologies… see HP SFS basically a bundled “storage grid that is Lustre ready, supporting 35+GB/s performance (ether) and 1PB of capacity built from small boxes (hp’s quote: ”3X the bandwidth at half the price of scalable NFS offerings.“)… Sun is trying the same thing with Constellation and their x4500 (thumpers), the challenge that we all have is that vertical scale brings systemness, but on a non-commodity price curve. The commodity/sedimentation / scale out architectures, however have a dirty little secret, as you scale out the complexity of management becomes high, and their provision becomes more rigid because of this… impacting capital arbitrage and schedulable density (fancy ways of saying, I have to lock apps to boxes to ensure that I can manage, instead of allowing apps to be dynamically scheduled to available resources (compute, network and storage).
Costperunitscale-2
In fact, what grid technologies do is add systemness to compute fabrics, and are hugely complimentary to existing virtualization / abstraction layers like VMware and Xen. So for me, Opsware is a ”platform“ that can help HP control the migration of their customers from static deployments/manual processes -> static deployments/automated processes -> dynamic deployments/automated processes.
I think that we’ll see other activity soon around BladeLogic from the like of BMC and IBM. The gauntlet has been thrown down!