Cloud Computing: a work in progress or a silver bullet?
I have seen some research recently that suggested the CIOs are not seriously thinking about Cloud Computing. Is this a leading indicator on this emerging market or is it looking in the rearview mirror? I vote for the rear view mirror theory. Here is why. If you are the typical CIO, you are thinking about everything from your budget, how to reduce energy costs in your data centers, proving to the CEO and CFO that your investments are indeed showing a return on investment or at least giving you a competitive weapon. In our research we are in fact seeing that many CIOs that are implementing new infrastructure and plans for new efficient data centers and innovative Service Oriented Architecture are making progress.
However, when I see research about doubts about clouds I can come to only one conclusion: fear of the unknown. What is a cloud? A cloud is an Internet based set of services based on a Software as a Service (SaaS) approach. Typically a single vendor controls this hardware, networking, software, and management environment.
Many CIOs and IT managers simply don’t understand what this means. It isn’t their fault. I have yet to see an article or announcement from a major vendor that makes it clear what a cloud really is (other than something that might mean rain). If I were a CIO struggling with all of the problems of a down market and requirements to make everyone happy I would be skeptical too.
Is Cloud Computing simply another word for outsourcing infrastructure? I believe that many CIOs will see it this way. After all, like outsourcing, clouds mean that computing is no longer on premise. There are obviously key differences between Cloud Computing today and outsourcing. The most obvious difference is the rationale for use. Today companies tend to use clouds for a specific test environment or to in essence host an application by a trusted supplier.
Over time, I think that CIOs will come around and accept that Cloud Computing is actually a valuable approach that is cost effective and trustworthy. However, in my view, it is going to take careful planning to gain the trust of business oriented CIOs. Here are what I think are the top challenges for achieving commercial clouds.
1. A cloud can hide many benefits and many sins. Therefore, a CIO has to be able to get under the hood of a cloud environment so that it is clear what the technology architecture is. Many problems are already surfacing because the existing cloud environment cannot scale and does not have a sophisticated management capability.
2. What exactly does the organization want to use a cloud for? Is it in place of a data center? If so, the CIO needs to do a lot of homework and establish a very well constructed Service Level Agreement with financial incentives and penalties.
3. Does the organization view a cloud as a standalone environment for one use? If so, how does it connect to an existing infrastructure and how easy is it to move data and other content from one site to another?
4. What happens if the cloud fails? Is there a back up plan? This is especially important if employees and/or partners and customers are dependent on the application or system that lives in the cloud. What happens if the cloud supplier goes out of business?
5. How proprietary is the cloud and how does that impact integration? This is important if you decide to leave one environment for another or even decide to bring the application back in-house. For example, some clouds may have designed their own languages for integration that might mean that an organization is stuck with an expensive rewrite.
Conclusion: there are no silver bullets or silver linings. Now, I think that Cloud Computing is going to be a very important transition in the maturation of distributed computing. In the long run, it will provide the type of utility computing that some of us have been talking about for decades. However, like anything else in the technology world, it is not a simple fix to complicated problems. It is an IT infrastructure made up of technology components that have to be managed, scaled, and secured – to name but a few issues. I expect that we will see a lot of failures in the coming year that will seed doubts among potential customers. At the same time, it will open opportunities for smart companies who have the vision to bring the pieces together and make this stage of computing a reality. At least we won’t be bored!