Hello,
I've been p2ping for a long time... I remember how proud I was of the whole napster ban (f*ck you metallica, one registry hack later and i was sharing all your crappy sellout albums
), edonkey/emule, bit torrent... as well as more 'server/client' style single p2p networks like winmx, direct connect, etc.In my mind there are two current p2p networks of viable size and support; emule/ed2k and Bit Torrent. Both have great similarities, they both have community sites based around the posting of verified hashes by users (ed2k has greater advantage of not being tied into one particular server of acquisition and the fact it is a simple text hash over a file, whilst bt has initial bootstrap gains from stating the tracker you must get it from.), although in my experience there's a lot more detail and quality associated with ed2k posts (fileheaven, etc. style verification)... maybe due to the much slower speeds of acquisition). There's also a similar hash and queue system, although bit torrents is much smaller which is why its speeds are far greater.
I prefer emule myself.... it is big, it is open, and it was revolutionary in its content/hash tracking. I love the community sites that sprang up from them, and have been involved in running several. However, I am aware of its age and from that age its inherant flaws, it is only a viable client if you leave it on a fast conneciton 24/7 and even then its speeds are off putting. edonkey was built around traditional connection speeds, and obviously its queuing mechanism is flawed because of it; its huge queues and overly large chunk sizes mean that it is, as the name suggests, slow and steady. In a modern era of fast connections, smaller chunks and smaller queues would be infinitely more beneficial as it vastly increases download time. This is the main benefit of the BT network, although there are scores of negatives, speed is greater. KAD has come along to combat the weakness of the emule network structure, although it does not address the flaws enough.
Torrent is much touted as the "best, coolest kid" on the block, it is indeed faster but it is deeply flawed and deeply insecure. The whole tracker problem ties you into what can only be described as a security nightmare; when one of these sites pops off the face of the planet and into an anti-piracy agencies mits they have a damning list of information, times, dates, ips, shared files, initial uploader, detailed share information, etc. all in clear text. The tracker mechanism is a curse much more than a blessing, and DHT adoption is hindered by "privacy"/clubs. Whilst ed2k is a wide open network, bit torrent is a splintered network by the tracker system. DHT has been a failure because of the splintering of the network, and its something that the creator and many observers continually criticise. Rightly so in my opinion. Its superiority in technology over ed2k comes from its smaller chunk sizes and tiny queues; you start downloading quicker. It is much more a single download at high speed than a volume network.
I myself, do not use torrent anymore, I have found that newsgroups and emule provide a better rate of consumption than emule and BT, or BT on its own. However, I am looking for a network that attacks the flaws of these two networks.
All networks suffer from the same flaws; upload is weak across all networks. This is largely a flaw created by the telephone companies and will continue to exist for a long time. This means that in an open network you are going to be let down by the fact that ultimately, nearly every peer can take out much more than they can put in. Ratios improved matters, but not by much. Splintering a network seems to help, but at the cost of availability and retention, and this is too big a cost. Trading and profiteering needs to die, and the splintering of a network is a risk to this.
I doubt any of this is revolutionary, and there's probably points of exception (please pick at them if you wish, my intention is only to highlight huge gaping holes
), however my real questions are the following:- What are the viable up and coming P2P projects that attack the risks of the big two above?
I believe security is going to be a big thing in the future, a tracker based network can't really viably persist. Similarly with recent attacks on emule servers, the open server model isn't looking bright either. So what is out there? I believe that hashing is an absolute necessity and the only way a network can exist now, communities increase the quality and quantity of content... they are a point of attack for negative forces, but they are an absolute necessity. - Ultimately how do you see the flaws of BT being corrected to make it a more viable network?
- Do you believe it would be in emule's best interests to redesign their core protocol and roll out a new network in their emule client to ween users from the older protocol? In my mind this has never once worked and just splinters and dies, for example with overnet (which in actuality was not a particularly revolutionary redesign).
) for sentimental value , just to grab the odd quality release they can't get elsewhere.
