MIME-Version: 1.0 Server: CERN/3.0 Date: Sunday, 01-Dec-96 19:13:15 GMT Content-Type: text/html Content-Length: 4857 Last-Modified: Friday, 03-May-96 19:03:40 GMT The Real Deal of the $500 Network Computer The Real Deal of the $500 Network Computer
by MULVA Inc.

Jeffrey M. Moore
jmm@cs
William W. Lee
wwlee@cs



...Check out the Mulva Report!

...And just what is this madness???

With all of the hype surrounding the new network computer and platform independent programming languages, such as Java, we decided to design our own network computer to see what type of hardware and software we would get using at the $500 price point. A $500 computer will not contain many things which we take for granted when discussing personal computers. For example, a hard drive becomes optional equipment, as does a keyboard and mouse. By removing these devices presents an interesting problem and raises many other questions. We want to find out what these problems are and obtain a solution for them. In order to do this, we must first design a network computer of our own to discover exactly what $500 buys.

We priced it using off the shelf, readily available hardware. Our implementation is only theoretical because we did not actually build a working prototype (mainly because computer scientists and that's not what we do). The main purpose of this exercise is to get an idea of the performance obtainable from a network computer. We want to determine how much of the network computer hype is real and how much of it is marketing. For more information, refer to the NC Specification document.

...and what do we learn from this?

A lot. We learn how to estimate the performance of a system that currently does not exist. This is a valuable skill because this type of evaluation must be done on a new project of this type to see if the end result will be usable. If the result product is not usable or the performance is less than that of what you can currently get, it will not sell. With this in mind, we want to know if the Network Computer will revolutionize the computer industry or just be another fad that will quickly die. Also, there are some other social aspects to this. For example, if the machine does not contain a hard drive and you receive electronic mail, where is the mail stored? This would be a new paradigm for the computer industry and we are trying to determine whether the NC is technologically superior to a personal computer or if it is a lot of hot air.

...and how does it tie in with High Performance Computer Architecture?

We trying to see if this computer is high performance at all. From what we can tell, we can't see that this type of computer provides at least the same performance as other computers in the same price range. It is not clear that this computer will perform at a speed which closely resembles a modern personal computer.

...and how do we measure success?

If we can get data that conclusively shows the performance of a NC, it will be possible to estimate the overall performance of the system. For example, if we use a NC with an ethernet connection browsing the WWW, will it have the same performance as a personal computer browsing the WWW. These are just some of the questions that must be answered.

...and finally, how are we progressing?




MULVA - Name of Jerry Seinfeld's girlfriend which rhymes with a part of the female anatomy.
Courtesy of the "Junior Mint" episode of
Seinfeld.

Visitors since April 17th, 1996