Instead, I like to imagine cloud computing as utility computing. As brilliantly explained in the book “Big Switch”, cloud computing is transforming the computing industry in very similar ways to the electricity industry in the last century. As electricity went from being produced locally on-site at mills, factories, etc. to centrally at a power station and then fed through a global grid, it became a utility that end-users came to expect and rely on. Similarly, cloud computing will manifest the desktop and office apps that we see now as centrally/remotely hosted and maintained apps that people will use as utilities.

I think eventually we will see enterprise-level services available to the end-users that are simple and truly plug-n-play. Similar to plugging in to the wall socket for electricity, end-users will be able to plug in to the next word processor, spreadsheet application, etc. and have it available at lowe cost without any limitations.

Thank you for your insight, I agree with it to a great extent. Easy, resizable, API-based deployment is indeed one of the biggest benefits that we’ve got out of the concept of cloud computing. But in my mind that is only an implication of cloud computing, not cloud computing in and itself. In fact, it is probably one of the more obvious outcomes. Cloud computing has the potential — definitely not yet fully realised — of going much beyond that. Storage provided as an API-based service, for example, is a radically new concept brought about by the advent of cloud computing and the semantics of it are so different that it is almost futile trying to compare it with any existing paradigm of computing.

The benefits of cloud computing could go well beyond server technologies — Desktop/OS for a personal computer in the cloud might even be the Holy Grail! Imagine this: your personal computer is just a dumb terminal that fetches the Operating System from the cloud at boot time and uses the cloud as permanent storage (with local caching for quicker access). This idea might seem far-fetched but it is already being used by some DVRs such as TiVo (only the fetching OS from the cloud part). Widespread availability of high bandwidth is the only major hurdle before we can realise this dream.

The proprietary nature of these systems was indeed discussed, although not for too long — everyone expressed their worries but nobody had anything more to say. The fear of lock-in is probably the biggest impediment to the widespread adoption of cloud computing. Developing using a standard framework and then having it deployed over the cloud is one way of avoiding such lock-in; the service offered by Heroku (Ruby on Rails Platform as a Service) is a good example of this.

10gen seems very interesting, I’m surprised that I had not come across it before. I believe such efforts are essential for the future of cloud computing — apart from the obvious benefit of providing a non-proprietary platform, the wider adoption of such systems will increase competition thus kindling innovation and bringing down prices for the consumer. I’ll surely download and try it out.