Today I was in Warsaw and had the opportunity to listen to Nicholas Carr, the author of “The Big Switch” describing how the provisioning of electricity changed over time, and how he believes IT is moving in the same direction. Will IT in general and compute power in particular, be centralized in the same way as large power plants generate our electricity. I’m not sure I agree fully with his analysis. First, cloud computing is a fuzzy concept addressing provisioning of both infrastructure and software.
Let’s start with infrastructure. That’s probably the concept that is closest to electricity. It’s pretty commoditized these days. And virtualization shields our programs from that infrastructure, serving as a “de-facto” standardization layer. A couple variables exist (e.g. Xen, KVM, VMWare), in the same way as with electricity (110 or 220 V, 50 or 60 Hz). It is clear that infrastructure is no longer a competitive advantage. From that perspective we could consider IaaS to be the great equalizer, giving the same capabilities to all enterprises at roughly the same price.
Somehow, in the back of my mind, I still remember that, 30 years ago when I entered the IT business, I was told the mainframe was dead. Well 30 years later, it’s still there. So, will cloud take over, and become the WWC, the worldwide computer? Or will it complement the mainframe and client/server environments? Nicholas Carr predicts it will take 20+ years for IT to disappear in its current form, and there I agree with him. IT departments are doomed to shrink. However the change may take much longer to be finalized.
To maintain the analogy with energy, the government had to establish clear rules to avoid monopolies, ensure competition and protect the user. The same will probably have to happen in cloud. In the current free market economy, that might be difficult to accept.
Software, or services as it is called in the cloud world, is a different ballgame all together. Indeed, it is the software that encapsulates the way companies are working. Does that mean they cannot be commoditized? That’s not what I said. Actually most software functionality, no longer provides competitive advantage. I remember a number of years ago, a chemical company taking the stage at a conference, pointing out they were the 32nd enterprise (out of 35) in their industry to implement SAP. So, having an ERP system is no longer an advantage, but not having one is a disadvantage. Using ERP functionality in a SaaS model makes a lot of sense. But that implies companies are ready to use the functionality as is, and no longer to customize it to their specific requirements. Actually CRM and Salesforce.com have proved it’s possible. Carr predicts HR is on the way and ERP will be next.
This brings an interesting point. Companies should identify their true competitive differentiation, and establish what IT functionality they need to maintain it. For some it may be product development, for others business intelligence or customer interaction. That’s where investments are required, where an extremely agile and flexible environment should be used (a private or public cloud infrastructure supporting a service oriented application environment). That’s the place were companies will move away from pre-canned applications. Even if cloud infrastructures are used here, they are there to support differentiation, not to be the great equalizer.
For everything else we should move to the cloud as soon as appropriate services (addressing the security, service levels and compliance concerns) are available. The implication of commodity is that things are getting as cheap as possible. The question IT departments will be faced with is whether they return the savings to the business, or take advantage of the savings to improve the differentiation in this small number of areas where IT helps the enterprise being unique. It will probably be some of both.
To come back to Nicholas Carr presentation, he highlighted “Cloud as a Revolution” amongst others. In his mind IT traditionally works in isolation. And I tend to agree with that. We have our own vocabulary, acronyms, nearly our own language. Dialogue between IT and the business is often difficult. This is why business requirements are poorly translated in IT in more than one occasion. IT looks at four key elements, infrastructure, applications, data and users, but often looks at them separately.
The business is more collaborative in nature. It’s all about sharing. He argues that the cloud (r)evolution forces IT to change, to make new choices, to take new roles. If we believe IT should become the strategic service broker to the businessIs Cloud , as I discussed in previous posts, collaboration with the business needs to improve drastically.
Whether cloud becomes the great equalizer or not, “cloud is a one way street” says Nicholas Carr. So we should expect an increasing portion of IT to move to the cloud (private in the short term, public later), and delivery to be centralized with a limited amount of service providers. It’s interesting though, that, after nearly 100 years of centralization, energy production is now going back the other way. The increased use of solar and wind, decentralizes the production again. Will the same happen to cloud, or will cloud bypass the central stage by developing a hybrid model? I’ll address that in the next blog entry…