Cloud computing: OS vendors adapt fast or wither away...

So the US gov has gone into the (private) cloud business (ref: 1 , 2).  With all the regulatory, security, complexity and scale issues inherent in the government space that a major gov (and there is none bigger in terms of annual IT spend than the US) can commit to making a cloud and making it work is remarkable.  Soon the complexity of building a cloud will lessen, it will become easier, and perhaps even a commodity to deploy in cloud units.

This is a huge opportunity for OS vendors. Whilst some got rich on "commodity computing" that was just for openers. The real commodity/utility computing is now:

- millions of of very cheap computers in the cloud
- millions of of very cheap computers in the hand (netbooks and cellphones)

both 100% require a cheap OS - there is 0% likelihood of demand for an expensive per server license.  The $1500 per year OS is yesterday's today that does not realise it is today's dinosaur.

There is a lesson to be learnt from Dell's desktop pricing model in the UK.  For years it absolutely refused to sell cheaper than £999.  Every new advance meant that the spec was upgraded but the price was the same.  Then the model broke and desktop prices came tumbling down.  I think the same thing will happen here to OS vendors with an expensive per server licensing model.   Far better to augment with very affordable per cloud unit (bunch of server) pricing.  Maybe keep that $1500 per year price but for 20 identical stripped down servers not 1.  Figure out what needs to be done to minimize server complexity and support costs, yet maximizing specific compute power.  Support customized stripped down kernels and OS footprint for specific services (web, db, load balancers, whatever) - and commoditize them.  Then pile them high and sell them low.  Otherwise the cloud providers will simply do that themselves and hire all your best support people to run them.