Migration, Integration and the Right Way to Build a Cloud

A single business unit spins up a cloud environment and starts using it for non-critical storage or batch processing. No big deal.

The enterprise as a whole implements a wide-ranging cloud strategy intended to provide rapidly scalable infrastructure and fully integrated data architectures. Big deal.

The open secret in the cloud is that while it does provide substantial improvements in data flexibility and infrastructure scalability, there’s more to it than just signing a service contract and letting the data flow. As I mentioned a few weeks ago, cloud migration continues to be one of the main inhibiting factors in getting greater enterprise buy-in of the cloud concept. Leading developers are issuing a steady stream of new platforms aimed at streamlining the migration process and integrating internal and external data architectures, but a fully functioning cloud environment requires more than just advanced software — it takes an entirely new enterprise mindset.

This realization has only recently begun to set in among top enterprise executives. As KPMG International discovered in a recent survey, about a third of cloud projects run over-budget because enterprises fail to factor in all the ways in which existing infrastructure and processes must change in order to achieve the desired results. These include reconfiguration of business processes, IT management, systems integration and security. KPMG concluded that coordination between IT and business management is crucial in order to identify and implement realistic goals for the cloud.

When it comes to the actual migration, there are three key elements that can make or break a project, according to Bluelock’s Jake Robinson. First and foremost is the selection of a migration strategy, which can range from simple data movement to virtual machine replication to full physical-to-cloud migration. Secondly, all data should be evaluated to determine its importance, or gravity, so as not to flood the cloud with non-relevant data or applications. And third, a proper assessment of the inter-relationships between applications is needed to ensure data flow and coordination is not hampered by the new cloud environment.

That last point is probably the most crucial aspect of the cloud, according to CloudVelocity’s Greg Ness. Simply migrating applications to the cloud makes no more sense than shuttling them to a co-location facility. Only through the smooth interaction between cloud and legacy infrastructure will enterprises see any real value in this new data paradigm. At the moment, full integration is something of a pipe dream, but at least the ball is rolling in that direction. And once it finally arrives, the enterprise will be open to levels of flexibility and dynamism that it can’t even imagine.

At the moment, the cloud represents the exciting new frontier in data infrastructure. But it won’t be long before the term becomes synonymous with garden-variety IT. There are those who even argue that internal physical infrastructure will soon seem as quaint as steam engines and penny-farthing bicycles.

If that comes to pass, the challenge will shift from migration and integration to ground-up design and implementation.