Xerox Corp’s Palo Alto Research Center has developed a system to improve networking in which one user’s computer bids for time on another’s computer. This development follows a long line of inventions from the Research Center which numbers among its successes the development of the networking medium Ethernet, as well as of the mouse cursor […]
Xerox Corp’s Palo Alto Research Center has developed a system to improve networking in which one user’s computer bids for time on another’s computer. This development follows a long line of inventions from the Research Center which numbers among its successes the development of the networking medium Ethernet, as well as of the mouse cursor driver and screen icons which Apple has taken to the core of its computer image. Xerox’s latest research project is, reports the New York Times, called Spawn, and was set up to explore ways of reducing traffic jams on computer networks which squander computer time and place the bother of co-ordinating computer access on the shoulders of frustrated users. Under Spawn, users requiring fast access to a particular machine can put in a high bid for entry to that machine, and by placing a high enough bid can even (in a very un British way) jump the queue of waiting users. Conversely, when there is only a small queue the bids placed can be quite small, and when there is no queue computer time is free to users. (Unfortunately no clues were given as to the currency involved in these transactions, although presumably a higher authority would in practice set priority levels on each of the activities being undertaken – or each user might have a set budget of computer time to use as he saw fit). The system will create a distributed computing environment in immense networks of multi-vendor machines, since it does not need a centralised controlling computer. This will have advantages in the future as networks are developed to such a huge size that the attempts at a centralised computer to control all the updated information on a distributed network will be thwarted. For the moment, however, Spawn has two main uses. Firstly, it can locate idle computers in a system and buy time on them for the user who requires capacity for a large computation, as well as offering users who may have had their computers hijacked for the aforementioned computation, time on a remote machine. Secondly, the project is being used as a workbench to simulate the behaviour of complex systems. It remains to be seen, however, whether, should Spawn grow into a success story, Xerox will make money out of its latest research venture, or whether the idea will be licensed fairly anonymously along with its mice and its icons and its local area networks, so that it is once again others that make their fortunes out of the latest of the brainchildren out of PARC.