Doing science by stealth
August 30, 2001 3:54 AM   Subscribe

Doing science by stealth Scientists have found a way of subverting the error checking mechanisms of web servers to allow them to perform calculations without the owners permission. This "Parasitic computing" could potentially use the internet as a single giant distributed computer.
posted by astro38 (5 comments total)
Unethical yes, but damn if that's not one of the cleverest things I've seen! Using standard TCP checksumming to perform distributed computing... man, I want some of whatever the person who thought that up was smoking!

Aside: I've long been fascinated by NP-class problems such as the "travelling salesman problem", but I thought the whole point of solving NP-class problems- where supposedly finding a specific solution to solving an NP problem could lead to a general solution to all NP-class problems- was to find a way to solve them short of brute-force computation, even if a brute-force method has been found.
posted by hincandenza at 4:34 AM on August 30, 2001

Slashdot thread on the topic.
posted by Mo Nickels at 6:14 AM on August 30, 2001

Yes, you could expand the pool of available computers for distributed computing, but ... right now there's far more participating in SETI@HOME than are actually needed.

Also, as noted, this is massively inefficient, even if it were ethical. It's almost like a DDOS in reverse.
posted by dhartung at 6:53 AM on August 30, 2001

I have to wonder about break-even. Is the compute expense of creating the packets and sending them greater than the compute cost of actually doing the work locally? It's not so much whether this works as whether it's actually worthwhile.
posted by Steven Den Beste at 9:00 AM on August 30, 2001

Kind of makes you wonder if it's already being done...
posted by fusinski at 2:04 PM on August 30, 2001

« Older Redbird reefs of the coast of Delaware   |   101 Dave Eggers Jokes, a.k.a. 101 Side-Splitting... Newer »

This thread has been archived and is closed to new comments