Private Clouds Brighten the Post-McKinsey Gloom

You have to give McKinsey & Co. credit — its report questioning the cost efficiency of cloud computing has legs. It has been more than two weeks since the report was released, and the hits just keep on coming. The report has raised the hackles of public cloud supporters (rightfully so), and it has raised the profile of private clouds, which McKinsey summarily dismissed when it suggested virtualization as the on-premise alternative to the cloud.

Ignoring the benefits associated with private clouds is particularly troublesome, because enterprise customers simply are not going to migrate en masse to public clouds any time soon. Large corporate customers are experimenting with public clouds and even moving some low-hanging-fruit workloads to the cloud (e.g., pharmaceutical research or Monte Carlo simulations), but most of the serious movement toward the cloud is being done internally. Given the money already invested in data centers, and the requirements attached to many enterprise applications, looking to internal clouds is the prudent thing to do. (Even Forrester thinks so).

It also is important to note that internal clouds are more than just virtualized data centers. Their function is not just to make application management more dynamic, but to provide a service-oriented platform in which users don’t have to worry about servers or capacity. They simply run their apps as services, knowing the only resource constraints will come from internally defined policies. Eucalyptus just went commercial with $5.5 million in funding for its EC2-emulating private cloud solution, but I assure you the company will have to add functionalities like automatic scaling (which is on Amazon’s roadmap) in order to be successful. Like public clouds, private clouds are about self-service and automation.

Apart from more questions about the validity of its method, McKinsey also caught flack from a certain Mountain View, Calif.-based Web company. Google took exception to the blanket admonition against using cloud computing, at least inasmuch as it includes Google’s fantastically designed cloud infrastructure and services (read the post, you’ll see). The main problem with Google’s approach to refuting McKinsey is that instead of defending cloud computing as a paradigm (the very paradigm that was questioned in the report), Google simply touts all the great things it is doing. Not only has this been viewed as a shot at VMware’s cloud efforts, I know some private cloud vendors who are none too happy about Google’s “we’re better than everyone else” approach. (I’ll keep their identities private until they decide to blog about it.) Google’s approach certainly has its strong points, but it’s not for everyone.

I’ll remain in the private cloud sphere to conclude, possibly saying goodbye to Cassatt. Cassatt has been trying to sell private clouds since before there were private clouds, and its power management capabilities are beyond anything else I have seen. The problem is that Cassatt was (and maybe still is) ahead of its time, and customers were hesitant about the sweeping cultural and operational changes that accompanied its cloud-like approach. As former Cassatt employee James Urquhart writes, it simply could not gain traction with it “go big or go home” approach. Cassatt tried to remedy this by breaking its Active Response product into pieces to accommodate incremental adoption, starting customers off with the power management piece, but it appears to have been too little too late. I’ve been told there are a couple of acquisition options still possible, so we shouldn’t finalize that eulogy just yet, but it’s sad that it has come to this point for this innovative company.

Question of the week

How does Google’s cloud infrastructure stack up to the competition?