How can botnets be prevented from becoming distributed computing clusters?

[Mk]

First,
THIS IS NOT A HACKING QUESTION, I am just interested in distributed computing.

I was wondering if anybody around might know what prevents a "botnet" from being something like a WAN beowulf cluster?

Thanks, :)

 

[Mk]

How is a project like BOINC and SETI@HOME, different from the way a botnet gives large amounts of regular home PCs tasks?

 

[PerennialII]

First,
THIS IS NOT A HACKING QUESTION, I am just interested in distributed computing.

I was wondering if anybody around might know what prevents a "botnet" from being something like a WAN beowulf cluster?

Thanks, :)

...actually the differences between a botnet and for example a HTC cluster (of which have most experience) are fairly small I'd say. In both (at least with a typical config) if a user's workstation is utilized they don't necessarily have much control over what the cluster software is doing on their computers and when (Ok, if use personal workstations as a part of a HTC cluster typically configure the cluster appls to run only when the station is idle or during nighttime etc.). Sure a botnet if understood in a typical context goes "somewhat" beyond. How similar is the coding underneath and what protocols they utilize ... hmmmm

 

[mgb_phys]

One difference is the communications in the cluster.
A cluster normally allows direct communications between nodes, sometimes this is even a very expensive high performance network like InfiniBand rather than just ethernet.
A SETI or botnet cannot easily communicate between machines because of firewalls and so all coms is to the master node.