Clients: thin or fat or something in between?
Part 2 of my chat with Intel's rising star senior vp Anand Chandrasekher: (Here is part 1)
I asked Mr Chandrasekher if the change in Intel's microprocessor development towards low-power consumption might also reflect a move to more of a thin-client PC design. After all, a thin client would use a lot less power than one that required a large, complex general purpose microprocessor.
And, a thin client could offer a very rich PC user experience when we have ubiquitous broadband, which we know we will have, sooner or later. The general purpose processing of a microprocessor could be done out in the cloud (also more power-efficient).
All you would need on the thin client is some chips that have the ability to quickly render graphics, video and audio bits, decode, encrypt, etc. And maybe the chips can also be optimized for AJAX processing?
Then you have a less expensive PC with all the user experience of a fat PC, and at a much lower price. This would help bridge the digital divide.
Mr Chandrasekher disagrees, and he set out his reasoning:
When computers were expensive it made sense to time-share the system. When bandwidth was expensive it made sense to have the computer at the edges of the network to perform local processing. Now that computing and bandwidth are equally cheap, the result will be neither fat nor thin but something in the middle.
We'll see if that happens. I think Mr Chandrasekher makes a good point in his analysis, but, cost of the client will determine what the market chooses. And if I can save a $100 by buying a computing device that is thinner, and yet I get just as rich of a computer experience as that of a fat client, then I will choose the less costly one.
If the processing is done out in the cloud, out in the internet, then a PC type desktop or notebook computer just needs DSP and graphics/multimedia type processors, and those are a lot less costly than general purpose microprocessors. I might not need any Intel chips at all.
ATI Technologies or Nvidia--the two dominant producers of graphics co-processors could easily step into this arena if Intel doesn't want it. Intel can concentrate on the server microprocessor market.
- - -
Please also see: Microsft shifts from PC centric view--what about Intel?