low bandwidth function
 

for those users who are either limited by bandwidth (modem low ratio ADSL etc) port blocked (like most budget DSL/Cable in UK) or Transfer restricted (X Gb pcm)

the option to Download as normal but to give something back in return, Holding Segments of Files (Segments of files are just data, so are they technically infringing laws and ISP AUPs)

for Hi bandwidth users that have been threatened with getting pulled for sharing files, the option of acting as supernodes, or partial caches (ie segmented file cache as above) A few high speed outbound connections connected to a few good hosts, allowing multiple Low bandwidth connections to DL partials, again this would be an option haven't studied the Gnutella protocol too closely, and there doesn't seem to be too much that would lean in this direction, but new standards appear all the time right (eg Multiple host DLs) I believe this would allow the network to grow, as there are an awful lot more Low bandwidth users out there, (Can't remember the thread I Lost it wahhh) but there was an excellent thread on the possibility of file swarming / relaying(proxying) for Modem/non sharers.

the High BW users could keep a list of most requested files, then divvy the segments out to multiple Modem users, obviously the modem user would have to agree to a disclaimer that unknown dater, of value no more than X Mb would be written to their hard disk in say 512kb chunks, and that the material may not be legal in thier country or acceptable under their ISPs AUP, but if these are segments that can only be recombined with Xolox(or any program once a multiple host DL standard has been agreed) then the files would literally be just meaningless junk.

I would love to see the modem users help this network, and the people who are in danger of getting thier connection pulled by ISPs these people have the bandwidth, they have the diskspace, and some lucky ones have both (G) if we can just get them to give some sort of agreement to the point of every little bit helps

(Sorry for the length of the post, and as I said I haven't read the protocol docs too closely)