Everywhere you go these days someone is
talking up the merits of cloud computing. But upon closer examination it
becomes pretty clear that cloud computing is pretty much little more than a
variation of the traditional hosted computing model.
Instead of paying for a
dedicated set of servers to run an application, the cloud computing model
allows customers to dynamically scale their hardware requirements to match the
needs of the application. If capacity demand rises, additional server and
storage resources will automatically be provisioned.
Customers like this
model because they don’t have to pay for hardware resources they are not using.
In particular, the model is attractive for running relatively mundane
activities such as providing disaster recovery services and for supporting new
applications when the customer is not sure how strategic they will actually
will become. In the latter case, it allows the customer to experiment with a
new application while limiting their hardware expense in case the application
fails to live up to expectations. On the other side of that same equation, if
the application succeeds beyond their imagination, they can easily scale up the
hardware needed to support it.
The ability to deliver
this level of IT flexibility is increasingly going to be a requirement for all
solution providers. The question that many of them will have to decide is
whether they want to build this capability themselves or resell the cloud
computing services of other companies. Already, we’ve seen startup companies
flush with venture capital money rush into this space. Some of the better known
startups in this space include Skytap, CohesiveFT and 3tera. But at the same time traditional hosting
companies such as IBM and Terremark have seized on the opportunity alongside Amazon,
Google and Dell, and the odds are better than good that in the wake of its
merger with EDS we’ll see Hewlett-Packard in this space as
well. In fact, executives at Terremark are reporting that they’ve already
received a number of overtures from solution providers in the channel about
reselling their cloud computing services.
The implications of
cloud computing models for solution providers are immense. On the downside,
cloud computing models threaten managed services models. The whole idea of
cloud computing is to move the compute resources into a hosted environment as
opposed to managing them remotely on the client’s site. The good news is that
the cost of managing compute resource locally inside a data center managed by
the solution provider is a lot lower than trying to manage resources located on
the client’s site.
The second major issue
is the opportunity that will arise as a federated cloud computing
model develops. There is no way that most customers are going to have all their
compute needs serviced by a single cloud. This creates an opportunity for
solution providers to serve as the general contractor of a set of cloud
computing services that are actually federated across multiple sources. How
solution providers develop a profitable business model around those services
will prove to be the ultimate challenge.
It’s important to
remember that not all compute resources are going to be delivered via a cloud
computing model. What we will ultimately see is a blended computing model where
resources are dynamically integrated across the on-premise assets of the
customer and the cloud. Managing the interactions among all those different
computing models is what in the end will create new opportunities for solution
providers.
But one thing is for
certain. In these uncertain economic times, interest in cloud computing is
going to be on the rise and that means solution providers need to start crafting
a strategy today if they want to stay relevant in 2009.
Commentary 
