un pic de istorie p2p

The Eventual Fall of Kazaa
December 4, 2005
Thomas Mennecke

In 1999, a small file-sharing application named Napster came into being. Virtually an overnight success, Napster amassed legions of loyal followers. Napster provided two key characteristics that helped it become a success and role model for future P2P networks.

The first characteristic was the immensity of its resources. With a peak poplation of 1.5 million users, there were few music files that weren’t obtainable. All other P2P networks that existed at the time, such as Gnutella, iMesh, and CuteMX, paled in comparison.

The second characteristic Napster maintained was its communal nature. Being a derivative of IRC, Napster provided an extensive community atmosphere through its bountiful chat rooms. Chat room dedicated to a wide variety of topics, issues and other ‘interesting’ genres gave users a feeling of home.

But an extended longevity was not meant to be. The RIAA (Recording Industry Association of America) took great exception with Napster’s existence. Representative of the major music labels and 90% of all music that is produced, the RIAA felt is business model was threatened. In 1999, the RIAA began its long crusade against Napster, which eventually led to its downfall in early 2001. Napster was finished, but P2P technology was just getting started.

During Napster’s reign, two important developments occurred; the advent of Gnutella and FastTrack. Gnutella and FastTrack represented a new breed of P2P networking. Gnutella was a fully decentralized network, as no centralized server existed to index files.

Instead of search requests being routed and indexed by a server, search requests were broadcast throughout the network. Although advanced, this model didn’t quite work out. The growing size of the Gnutella network made it difficult for a search request to make its way through every peer, which caused network tie downs, traffic jams, and mismanaged bandwidth.

The solution to this was FastTrack. Neither fully centralized like Napster nor fully decentralized like Gnutella, FastTrack utilized something called supernodes (or leaf nodes, super peers, etc.) Supernodes are indexing servers that exist throughout the network. Instead of being centrally located like Napster, they were integrated into the Kazaa software. Any member of the FastTrack network could be a supernode as long as they had the bandwidth and computer resources.

This model would eventually become the standard of P2P networking, and various iterations are used by many modern P2P networks such as Ares Galaxy, ManolitoP2P, and interestingly enough, Gnutella.

Over time, FastTrack became the dominant P2P network. At its height in May of 2003, FastTracks poplation reached 4.5 million simultaneous users — almost triple the size of Napster.

Since that time something remarkable has happened. FastTrack’s poplation has plummeted. Unfortunately, estimating the exact poplation of various P2P networks has become more difficult in recent years due to false servers, flooding and architecture changes. Yet it is clear the FastTrack network is nothing like it once was. A recent study suggested that nearly half of all files available on this network were corrupt in some way.

FastTrack’s poplation decline and the corruption of files are directly related. For the last several years, various companies such as OverPeer flood networks with corrupt files. FastTrack is their favorite target, because it does not contain the same protection strengths as Ares Galaxy, eDoneyk2000 or LimeWire. In order to protect a P2P network from corrupt files, a hash code is associated with an indexed file. These hash codes are unique fingerprints that can be associated with a verified file. If the hash codes match up, more than likely the file is legitimate.

FastTrack using a very weak hash algorithm called UUHash. UUHash’s advantage is that is can hash files very quickly, but its major drawback is that it only hashes a fraction of the file.

Since UUHash hashes a file at predetermined integers, whoever wishes to corrupt a network with false files only has to make sure the proper integers are hashed according to correct standards, while the rest of the file can be polluted.

An accumulation of problems, such as the corruption of files, has greatly weakened the FastTrack network. Kazaa’s reputation dwindled as the inclusion of irritating and system degrading third party software took its toll.

The company at the helm of Kazaa, Sharman Networks, hasn’t released a significant upgrade since its takeover in 2002. The FastTrack network architecture remains unchanged, the hashing problems still exists, and corrupt files are still the hallmark of this community. One way or another, this will all change.

On December 5, Sharman Networks is scheduled to upgrade their key word filter to include 3,000 copyrighted works of the ARIA’s (Australian Recording Industry Association) choosing. The ARIA can update this list every two weeks with additional copyrighted works to reflect the latest pop-culture trends. The effectiveness of this filter will serve more to raise the curiosity of the FastTrack poplation rather than deter any kind of copyright infringement.

Many Kazaa users live in a bit of a dream world compared to rest of the file-sharing poplation. They are aware file-sharing exist, but only through Kazaa. Interestingly, many are completely unaware of the RIAA lawsuits, the lawsuits against P2P developers or Kazaa’s eventual transition to a pay P2P model. But a time of enlightenment will come.

With the delay in the Kazaa trial, thanks to the ARIA, Sharman Networks now has until sometime in February to develop an efficient filter capable of seriously deterring copyright infringement. That gives the FastTrack poplation 3 months of valuable time to abandon Kazaa and seek shelter in more reliable networks such as Gnutella, BitTorrent or eDonkey2000.

The next several months will prove to be a dynamic era in the file-sharing community. In all likelihood, FastTrack and Kazaa’s long reign will finally come to an end. The ARIA had a good opportunity to seize the moment and catch the FastTrack poplation by surprise. Considering many FastTrack users are uneducated regarding other P2P communities, a serious number could have been converted to paying customers. However with a slow migration to a pay file-sharing model, it’s likely the eventual fall of FastTrack will only help create awareness for other P2P communities. The fall of FastTrack and Kazaa has been a long and drawn out affair. Already half of its poplation has moved on. The next few months will only serve to hasten the migration of its remnants.

from here