How about gnutella servants implement a mechanism that automatically downloads the most popular files? This automatic download mechanism would start only when the computer is at idle and there is plenty of ample bandwidth. It would store the popular files in 1 large file, using x% of the hard drive’s free space.

The popular files wouldn’t necessarily have to be stored in whole. Storage could work like a striped RAID. Only sections of a file could be stored. This could make copyright infringement less of an issue.

Supernodes may be able to collect the top query hits and collect popular file information. Checksums would be used extensively to test data integrity.

I haven't fully researched this, but Freenet may have a similar mechanism? I would like to see something like this implemented w/in gnutella

Thoughts or ideas on this? Has this already come up?

mark

Hmmm--I don't know if I want any files on my computer that others find popular or that I don't know about--especially if I can't examine them. Interesting idea though (first I read of it, but that doesn't say much).

Yes,

Moak mentioned this almost two years ago. The original idea is from a smaller net and the technique is called "swarming", many people call the Gnutella multiple source download technique also swarming.

Morgwen