On the Hunt for ‘Elastic Computing’

The enterprise is quickly transitioning from the concept and design phase of cloud development to a deployment and operations phase. In many ways, this is where the rubber meets the road in that we will soon learn what the cloud can actually do as opposed to what has been promised.

Part of this proof-of-concept period is the idea of elastic computing. Can a cloud environment actually provide the kind of dynamic infrastructure that will allow users to pick and choose the operating environments, applications and data sets to achieve their self-defined goals?

More and more, the answer to that questions seems to be yes, although it will take some fairly sophisticated software to make it happen.

One novel approach comes from Translattice Inc. The company has released the new Translattice Application Platform that seeks to federate data center infrastructure across physical, virtual and cloud platforms to achieve the kind of elasticity that  data center executives have been calling for. The platforms relies on geographical distributed nodes, either physical devices or cloud instances, that house application deployment tools like load balancers, runtime data and storage. The nodes are tied to a globally distributed relational database that enables broad scalability and availability throughout the enterprise. The company says it provides a simplified and more fluid environment than traditional layered infrastructure.

Not everyone is willing to give up on layered management, however. EMC, for one, sees a simplified management stack consisting of three layers: a cloud-based information infrastructure, a composite application services layer and a user interface. In this way, the company hopes to provide a public/private cloud services environment, a robust development toolkit for content-enabled apps and universal, device-tailored, secure access platform for both developers and end users. To get there, though, enterprises will have to embrace practices like user-selected computing models and the extension of data governance into the cloud.

Data elasticity can be quite exhilarating, according to Kaavo CEO Jamal Mazhar, but it can also be intimidating. Data management through tools like Kaazo’s IMOD system (Infrastructure and Middleware on Demand) introduce levels of management that cut new ground across entrenched systems and data relationships. But these kinds of changes will take place regardless of whether you are aiming for simple consolidation or a fully functioning virtual/cloud infrastructure.

None of this is going to happen, however, without tighter integration between server, data, network, storage and applications management, according to Mark Shoemaker and Venkat Deraj of HP’s Software & Solutions Group. By shifting the focus from management of disparate resource and information silos toward a broader Business Service Automation (BSA) approach, enterprises will be able to avoid much of the cost and complexity of advanced information management in the cloud.

In a way, each of these perspectives center around the same basic concept: break down the barriers between resources, applications, data and users, and enable the smooth flow of information to wherever it can do the most good.

Such a scenario is certainly possible, but it will require the shedding of some long-held beliefs and practices.