Just about every aspect of enterprise software can now by purchases as a service, so it stands to reason that middleware would become available as a service as well. For example, we've already seen the creation of startup companies such as Linxter that offers a middleware service and the JBoss division of Red Hat has partnered with Amazon to provide a similar capability.

But unlike other software-as-a-service offerings, the prospect of having middleware available as a service has some profound implications for the future of enterprise computing.

Beyond the fact that developers can now rely on infrastructure in the cloud to create applications, middleware as a service will eventually turn the Internet into a mega-service bus for enterprise applications. The potential this creates in terms of integrating business applications across multiple companies is enormous.

Let's face it. Implementing, managing and deploying middleware software to integrate applications is still a painful, labor-intensive process. It's hard enough to do within an enterprise, never mind trying to do it across multiple companies.

And yet, the real value of enterprise computing in the future is going to depend heavily on how well it is used to support business processes that span multiple companies. To make that easier to happen, many IT organizations would be better served by relying on a standard middleware service they could simply invoke rather than having to build and manage their own equivalent.

There are, of course, performance implications for such services. A lot of mission-critical applications may not find the performance capabilities of such a service to be sufficient. But for a lot of applications where performance is not as critical, such a service could be just the ticket. In fact, we could see an explosion in business-to-business applications because one of the great inhibitors of creating these applications has been the cost and complexity associated with the underlying middleware.

When coupled with virtualization technologies, these middleware services have the potential to become a highway over which workloads dynamically move around a global Internet. For instance, an application running locally on premise in the United States from say 8:00am to 6:00pm could be moved across the Internet to another set of servers in Asia to create a dynamic IT infrastructure that was truly global while the middleware services kept all the integration points to that application consistent.

Taken together, these technologies could introduce a much higher level of flexibility into the underlying IT infrastructure that would be much better able to support the needs of rapidly changing business conditions. In essence, what is required now more than ever is the ability to manage and support dynamic application workloads that can be changed at a moments notice in order to support the business.

This would go a long way to bridging the historic divide between business executives and the IT community because when you peel back all the emotional arguments, what you find is the simple fact that enterprise IT today is not flexible enough to support business conditions that now change on a daily basis.

Naturally, dynamic computing environments where there is a lot of movement of application workloads are going to be a lot more volatile from a management perspective. But the management tools needed to support these types of environments will get better in time. And when they do, the first IT organizations that master these new platforms for enterprise computing are going to have a substantial advantage over rivals too encumbered to keep up.