That brings back nostalgia. I remember vividly being active on various eMule mod forums back in ~2005. There was a very active scene back then with some eMule Mods that really pushed things further, MorphXT, Xtreme, StulleMule, eFmod, Sivka, ionix, NeoMule and ScarAngel come to mind.
So much drama with leecher mods (leecher was negatively connotated in regards to the ed2k-protocol) that tried to download more than uploading or upload to clients only that had proven to pay back with higher speeds. But also really innovative ideas like downloading from clients that used a thin HTTP server within the client, leveraging your ISPs proxies (Webcache/Peercache feature) for higher download speeds. Also so many discussions where the network should be heading with the main Client devs famously mute in regards to those discussions and drama.
The main benefits of eMule/ed2K, compared to torrenting, were file discoverability (serverless thanks to Kad search) and longetivity of files. It was actually relatively hard to find a file that was dead, because people kept sharing as long as they downloaded, often much longer. Download speeds were much slower than Torrents, even back then, and that was an advantage because that kept so many files alive. Great times that basically ended because of „p2p sheriff“ companies, Torrents and the dawn of One Click Hosters like Rapidshare that promised privacy and better download speeds.
To this day I am wondering if eMule would still have an active userbase if it had some sort of mechanism that (via opt-in) allowed clients to download from trustworthy „friends“ only and pipelined those downloads from the „public“ ed2k network through these friends and their friends' networks. Yeah, may be a bad idea but to my knowledge never really got explored.
I loved, with eMule and/or Kazaa (don't remember anymore which one), the feature of being able to get the list of all what the remote host was sharing => my idea was that if that person was sharing something that I liked, then that person might have other interesting stuff which I might have liked => from time to time that was the case.
1 or 2 years ago I installed it, connected to the network (nodes? meganodes? forgot the terminology), and activated somewhere the parameter to be verbose about the incoming search queries => A LOT of nasty porn stuff being searched (I therefore quickly shut down and uninstalled everything, brr) => it's a pity, but thinking about that later maybe they were just automated queries or similar submitted by police & Co. to search for illegal videos, who knows.
My trick back in those days was to make an inverted filter. Instead of trying to blacklist the sheriffs, I white-listed peers.
Back then I was with an ISP that had an unusually large number of "peering" connections that were unmetered -- they wouldn't count towards the monthly download quota.
I tracked down the IP subnets for each peering company and then wrote a little program that would block everything except peered network ranges. This worked surprisingly well! Despite filtering out 95% of the Internet, my connections were short-range and high bandwidth, so I actually had slightly better download speeds -- and no more scary emails.
I did feel bad about one thing though: I couldn't figure out how to invert the subnet list efficiently, so I made a 4-billion entry bit vector (just 512MB!), flipped the desired ranges, and then exported from that.
I can feel my CompSci University professors shaking heir heads in shame.
>To this day I am wondering if eMule would still have an active userbase if it had some sort of mechanism that (via opt-in) allowed clients to download from trustworthy „friends“ only and pipelined those downloads from the „public“ ed2k network through these friends and their friends' networks. Yeah, may be a bad idea but to my knowledge never really got explored.
So much drama with leecher mods (leecher was negatively connotated in regards to the ed2k-protocol) that tried to download more than uploading or upload to clients only that had proven to pay back with higher speeds. But also really innovative ideas like downloading from clients that used a thin HTTP server within the client, leveraging your ISPs proxies (Webcache/Peercache feature) for higher download speeds. Also so many discussions where the network should be heading with the main Client devs famously mute in regards to those discussions and drama.
The main benefits of eMule/ed2K, compared to torrenting, were file discoverability (serverless thanks to Kad search) and longetivity of files. It was actually relatively hard to find a file that was dead, because people kept sharing as long as they downloaded, often much longer. Download speeds were much slower than Torrents, even back then, and that was an advantage because that kept so many files alive. Great times that basically ended because of „p2p sheriff“ companies, Torrents and the dawn of One Click Hosters like Rapidshare that promised privacy and better download speeds.
To this day I am wondering if eMule would still have an active userbase if it had some sort of mechanism that (via opt-in) allowed clients to download from trustworthy „friends“ only and pipelined those downloads from the „public“ ed2k network through these friends and their friends' networks. Yeah, may be a bad idea but to my knowledge never really got explored.