P2P Networks Classifieds

Three types of P2P networks are collaborative computing, messaging and affinity groups

Thursday, September 12, 2002

The P2P space can be classified into three categories based on the anatomy of the network and application. These are collaborative computing, instant messaging and affinity groups. Collaborative computing is also referred to as distributed computing and its biggest and most successful example is the Seti@Home project. Another is Fight AIDS at Home from Entropia. We all know about instant messaging and regularly use MSN, ICQ and Yahoo instant messengers. The hottest P2P applications in use today are the affinity groups like (erstwhile) Napster, Kaaza and Gnutella file-sharing networks. They are extensively used for music, movies and software sharing. Then there are the less well known workspace-sharing applications, like Groove, which also fall in this category. 

Kazaa is one of the hottest P2P apps in use today

While on one side, P2P technology puts unused and idle resources to constructive use, problems abound about the usage of such a network to trade software and multimedia content. In a corporate environment, the problem with implementing it is the lack of central control. Plus, there can also be security issues when individual desktops directly connect to each other. 

These debates apart, let us look at the P2P applications as they exist now.

Collaboration 
Seti@Home
Though the founding document for SETI was first published in 1959, it was not until the Internet got established as an effective medium to connect computers, that the Seti@ Home project came into being. Again, though the idea of using idle computing power of individual computers around the globe originated in 1996, it wasn’t until May 1999, that the software actually started its work. As of this day, it has around 3.8 million users and they have contributed a total of a little over a million years of CPU time to the project! The SETI@ home is yet to find intelligent life elsewhere in the universe. But the enormous amount of computing power expended on the project would not have been possible had it not been for this collaborative computing effort.

You, too, can participate in the effort. The only thing you have to do is to install the client (available for different platforms and OSs) and run it. The client automatically downloads data units, processes them in the idle CPU time and sends them back to the project base at University of Berkeley.

Fight Aids at Home
A similar distributed-computing project is Fight AIDS at Home (FAAH) from Entropia. Entropia is also running two more grid-computing projects—SaferMarkets and a non-profit research—but FAAH accounts for 80 percent of the whole grid. The FAAH project has a little over 50,000 machines running across the globe.

The Great Internet Mersenne Prime Search
Another similar project is The Great Internet Mersenne Prime Search (GIMPS). This runs computations for one of the most coveted finds in all of mathematics—a new Mersenne prime number. The task is again very CPU intensive. This computing community was also created by Entropia in 1997 and is hosted by the Internet PrimeNet Server (Entropia’s). The total computing power of the grid is currently equivalent to 69 of Cray’s most powerful T932 supercomputers! These distributed computing clients are also available for download and work in a similar manner as Seti@Home.

Messaging
Instant messaging is another widely used P2P application—MSN, AOL, ICQ, Yahoo and loads of others—that we use everyday. These messengers work in coordination with a central server. The server is, however, used to perform the task of authenticating a client to the network, storing buddy lists and making the first contact between two clients. Once a connection is established, two client applications talk to each other directly, though you also have the option of using the server to send messages. Instant messengers come with a lot of features that we overlook. We think of them as only text-messaging windows.

Take MSN Messenger. With it you can do message conferencing, application sharing, white boarding, transfer of files. It even allows you to make phone calls. Application sharing allows both interacting parties to work on an appli- cation together. White-boarding works like a virtual white board on which the interacting parties can write or draw.

There are also private messengers. They consist of a server and a client part that a corporate can buy and install for its network. This can then be used for similar functions as public messengers within the organization. These private messengers offer better security and encryption features, apart from local manageability.

Affinity groups
Popularized by file-swapping programs, this area of P2P consists of lots of applications, the most popular being Kazaa and Napster. Though propounded by their creators as ‘efficient usage of your available bandwidth’, these are sadly turning out to be platforms for illegal file swapping. Moreover, they are becoming bandwidth hogs. Music companies have been on the forefront of the trying to close these file-sharing networks down. The latest in the field to dissuade users from illegitimate music file downloads is a company by the name of OverPeer. They upload dummy music files to these P2P networks (files with a chord or chorus in a loop instead of a track).

But fileswapping is not what affinity groups are all about. Collaborative groups (also called P2P Groupware) can be public or private corporate groups that work on a common application. Two such well-known applications are Microsoft NetMeeting and Groove. These are what are called workspace-sharing applications. In NetMeeting, for instance, individual users can host conferences and invite others to them. Users in a conference can share their applications. Other users get windows that have visualizations of the shared applications. Groove uses a slightly different approach. Here you have workspaces that are created and shared among participating users. It uses a client-server mechanism to find peers and initiate services. After that, all communication happens directly between the participating peers.

Ashish Sharma

Page(s)   1