As could be observed with a reading of the comments on Mikes posts and Slashdot, there is strong opposition to the idea of centralising computing. I estimate that about two-thirds of the commentators opposed the idea. What were the arguments against centralised computing?
The three key criticisms (all somewhat related) were: lack of user control, concerns about privacy and confidentiality of data, and dependency on the telco or service provider. Let us consider each of these points.
In a centralised computing model, users have very little control on what they can do both on the access device and on the applications that are available to them. Some of these can be addressed by smarter technology, but there is a need for service provider concurrence. This is not very different from the world of mobile phones where the operator has significant control on what users see and do with their mobile phones. This is very different from the world of computers and the Internet as we know it now. In addition, since all of the users data will be stored on central servers, there is always the danger that something may go wrong (system crash, hacking, operator bankruptcy) which could make the user data inaccessible or, even available to everyone.
Localised (personal) computing allows users to customise their experience. Model T is not everyones idea of a car! Users can attach the peripherals they want and run the software they want without a perceived big brother having to approve. In a way, centralising computing is like asking car owners to give up their cars and start using public transportation. However good the public transport system is, people used to the freedom of their own cars are unlikely to switch.
Among other criticisms, there is the issue of latency the reaction time. No network connectivity can match the speeds of local disk access. Besides, for specific uses of the computer like gaming (which is very popular), the network response times may just not be good enough. After all, there is no such thing as infinite bandwidth. In addition, there is also a belief that people may be more comfortable with one-off payments rather than the thought of paying on a subscription basis.
And finally, there is the issue of history. Various previous efforts like the network computer and WebTV have failed. Users have spoken, say the critics of centralised computing. Why then do we revisit the issue? In tomorrows world, processing power and storage costs are available in plenty so why bother with anything other than the personal computer which has served us so well for more than two decades?
The answer to this lies not in the current users but in todays non-users. To understand the future of centralised computing, we need to think not of the first 700 million users of computers, but the next billion users.
Tomorrow: Utility Computing in Emerging Markets
TECH TALK Microsoft, Bandwidth and Centralised Computing+T