Forget Cloud Computing, What About Cloud Apps?
So, I have to admit, the combination of iTunes and Shazam is not all that good for my bank balance. But, outside of helping me fill out my iTunes library with songs I didn’t know I wanted, Shazam, along with the Google iPhone app, highlight three interesting trends.The first is that they usher in a model where information comes to you, which is 180-degree shift from the way things used to be. In the old days, you would dutifully assemble your pile of punch cards, walk down to the data processing department and drop them off, then come back later and pick up your printout–quaint, huh?. If needed to research, say, the airspeed velocity of an unladen swallow, you would have to go the the library. Today, the info comes to you, wherever you are. This decoupling of information creation, information repository, information processing and information consumption will have huge repercussions for infrastructure–I think we are just scratching the surface right now. Today, we are happy if we can gain some level of workload portability, but I think the things that will trip us up as we try to operationalize this outside of the data center are the second and third order dependencies for those workloads. Discussions to date seem to treat workloads as these discrete entities, I think largely because of the Vmotion model, but the reality is that the edges of the workload are seldom that clean.The second trend is the idea that a cloud or web based app will deliver greater performance and features than desktop-based equivalents. The reality is it is unlikely you could run (or would want to run) apps like Shazam or the Google app on you laptop because of things like processing, storage, etc. However, the initial discussion of moving apps to the web (Google Apps, Office, iWork) is around delivering near-desktop performance at a better price point or lower TCO. While appealing, especially in this economy, I think, at some point, that approach has limited appeal. Instead, I think the true tipping point is to deliver feature sets, performance, etc on cloud based apps that are not feasible on the desktop or the laptop. For example, I would happily use a cloud-based version of Final Cut if rendering was automatically tied to some sort of elastic processing capability.Finally, I think these types of apps are another nail in the coffin of the traditional PC form factor. These types of apps continue to reduce the hardware requirements on the client side to the point we can completely dis-intermediate the PC altogether and cloud intelligence can be embedded directly into the everyday devices people use day-in and day-out.