If the Cloud Only Had a Brain

One of the challenges associated with cloud computing is the huge scale of the endeavor. It’s relatively simple enough to centralize all the compute resources, but optimizing cloud computing deployments to behave as efficiently as possible is a whole other adventure.

That’s the challenge that companies such as Adaptive Computing, Platform Computing and Elastra are gearing up to take on. The basic idea is that as virtualization advances, application workload as going to be dynamically moving across both public and private cloud services. Managing that will require a high degree of automation because throwing people at the problem isn’t going to be responsive enough or economically feasible.

Adaptive Computing and Platform Computing have approaches to the problem centered around management frameworks originally developed for scientific applications running on supercomputers that are typically networked together in a grid computing framework. Elastra is a startup that provides a management framework that was developed by some of the pioneers of cloud computing services.

According to Peter ffoulkes, Adaptive Computing is pretty far along in delivering on this vision in the form of a Moab management suite that automatically provisions physical and virtual servers, manages workflow, dynamically distributes application workloads, provides a policy engine and enforces service level agreement. In addition, Adaptive Computing has marketing and sales agreements with Hewlett-Packard and IBM in place for its software.

Obviously, every company from VMware, making virtual machine software, to all the providers of applications and systems management software, has cloud computing management ambitions. The only real question is whether they will build out these capabilities themselves or move to acquire companies that already have.

In the meantime, chief technologists exploring the ins and outs of cloud computing might do well to take a step back from the underlying infrastructure in favor of a higher-level approach that focuses on how to manage the cloud first and what goes into it second.