economics of distributed computing
posted by windley, this is an interesting idea of breaking down the economics of computing on a very large scale. the idea that sneakernet (passing information via real-world transports like UPS or your own pair of sneakers and a floppy) is the most cost-effective to move a TB of data is pretty interesting. i would like to think that is simply a limitation of our networks that will be handled in the near future.
but the point of the discussion is the idea of breaking down the services of computing into parts like storage, bandwidth, database accesses, etc. much like other industries outsource (see article for examples like ATMs).
what if there were thousands of companies out there that needed an infrastructure for massive web services, but didn't want to be in that business? would others step forward and provide the service of providing services? would there be a few companies that compete in the Data Storage Market? would they be necessarily tied to the Bandwidth Market? what about services like that of Inktomi? will those disappear in the future of super networks? im thinking "probably not."
when one takes a step back from these technologies and looks at them from a business/economic viewpoint, new realms appear. i don't normally look at things from that vantage, so it's nice to see now and then.
Posted by Steve on April 22, 2004 10:53 AM