The Theory of Leech Computing

Phil Frisbie, Jr. writes "I am defining Leech Computing as 'a program running on a client computer without user knowledge that can process data and report back the results, but otherwise does not effect the usability of the client computer and makes no changes to the client'. Leech Computing, Part 1 covers basic theory."

Comments

Idea

[autocracy]

Can we use this to create a distributed webserver that where each person who visits the site will serve copies of it? This guy's system can definitely use it! SLASHDOTTED

Wrong, but not a bad thing

[SirSlud]

> how could you possibly get data back to the server without the user knowing it

He says refresh and 'tricking' the user are the only ways (on form submits.) Wrong.

dynamic.php:

<script>
data data data
do do
calc calc
var me = answer;
document.write("&ltscript src='http://myserver.com/donate.js?answer=" + encode(me) + "'></scr"+"ipt>");
&lt\script>

That sends some data to the client, does some client side cals, and sends the data back to my server (although I have to respect the max limit of data one can send via form posts, but its the same with his more obvious methods.)

This is done all the time to count impressions in the advertising world. In fact, in a sense, advertising tracking online is already leech computing in some implementations.

BTW, the .. " as a close of the top level script tag.

Difficulties with the theory

[toby360]

I'm going to assume that for now the author goes for something using Javascript:
#1. Javascript is extremely slow. It's also interpreted, not compiled. Code optimized and compiled for a system can be a hundred times faster.
#2. Coding anything usuable for this type of application would require a good bit of code to be sent via javascript.
#3. The amount of processing it would take to:

A) Generate the web page to send to the user with the appropriate Javacode + whatever the user needs to process
B) User's computer to interpret the Javascript, execute the code, send back to the main host computer
C)Host computer recieves the data, decides where to store it, what to do with it etc.

And for the code to run and NOT affect the user significantly (meaning the processing done wouldn't be very much at all), all in all would likely require far more processing than it would if it were compiled on a server just running by istelf.

All in all it would be very inefficient, and probably faster for the server managing the data and generating the pages to process this information on its own.