Searching the Clouds for Value

The hype about cloud computing services continues relentlessly. Not only could IT architectures revert back to more-centralised systems, but completely new business models may emerge to disrupt the market. Yet I’m sensing the first signs of “cloud fatigue” are creeping in with the clients I’m talking to. I’m sure this is partly caused by the flurry of definitions and propositions that suppliers throw at their customers. But equally, very few people have figured out how to make (more) money with cloud offerings.

The questions that come up time and time again are: “When will customers start pushing enterprise-critical applications into the cloud” and “How big’s the addressable market?” For me, good yardsticks are the progress in implementing service-oriented architectures, and the adoption levels of software-as-a-service, as both are constituents of the cloud. Not surprisingly, I take a relatively conservative view. The industry has reacted to parts of the nebulous outlook (or just plain confusion) by focusing on private clouds, which lets them address particular security concerns and worries about interoperability. I’m tempted to call this old-fashioned hosting that uses new virtualisation technologies, but perhaps this isn’t sexy enough for marketers. Nevertheless, it underlines that the next big step is combining applications and services with existing cloud or virtualised infrastructure offerings.

Talking to investors, I’m asked who the winners will be. Most cloud offerings are commodity plays. Amazon, a pioneer in cloud computing, is the best example of that. But Amazon has no legacy business that it needs to protect. Conversely, for a system integrator or telco that outsources networks or infrastructures the stakes are high. Equally, software companies like Microsoft could see their sources of revenue cannibalised. So where will be the value of cloud offerings come from? If I knew all the answers, I could make a lot of money. Aggregation? Integration? Guaranteed service-level agreements? For the plethora of innovative start-ups like 3tera, Zimory and Rightscale, the business model might be to become acquisition targets rather than disruptive players.

For mobile operators the cloud scenario is by no means just a question of definitions. Even if people eventually stop calling Web 2.0 applications like Facebook “cloud” offerings, the underlying problem is how to cope with the exponential explosion in data. CCS Insight has written extensively about how more and more networks will become choked by the uptake of mobile broadband (see Shaun’s post here, for example. And mobile devices will increasingly become integrated parts of cloud offerings as the access to information becomes ever more pervasive.

So where will the value for operators come from? From an accelerated move towards services? From a strategic shift toward a sophisticated “dumb pipe” where quality of service might incur a premium? Or from completely new business models that we haven’t yet accounted for? Operators such as AT&T and Deutsche Telekom are trying to push cloud services through their system integration divisions, but this will be not enough to compensate for the margin erosion in their core business. Verizon Wireless is suggesting it might revert to a usage-based model for its LTE network and allow multiple authenticated devices. Should this happen, it would change the current market dynamics completely.

While writing this post from a hotel room in New York, I struggled to get a mundane thing like e-mail working with the local hot spot. Which serves (at least for me) as a reminder that there are many challenges the industry needs to overcome before IT services really become a utility like water or electricity. While I don’t have all the answers about who will be the winners, I wouldn’t bet too much money that many cloud services (apart from infrastructure) will be mainstream within the next five years. Yet planning for multiple scenarios is an inescapable necessity. My money would be on the co-existence of different delivery models, not least because of the many legacy applications. Or in other words: the forecast remains unsettled.