Peer-to-Peer: Toward a Definition

P2P Home

We are at the threshold of applying another paradigm to network and Internet architecture and communications. Many important industries regard this as a revolution and characterize it as the instrument for the Third Generation of Internet.

I am referring to Peer-to-Peer networks and their internetworks.

The concept of peer to peer (no hyphens) communications is not new. In its simplistic form it is usually structured as one-to-one through an exchange system. The common telephone connections use this model. Currently, a usual telephone communication requires a central exchange or set of central exchanges to make and maintain the connection between dumb terminals. This is not why Peer-to-Peer is regarded as the new paradigm architecture for Internet.

Smart terminals introduce the possibility of other kinds of communication architectures. A number of these have been implemented in networks and Internet. Among these new possibilities, Peer-to-Peer (P2P) has been introduced showing a self-structuring scalability hitherto unknown.

Three important characteristics of Peer-to-Peer group members.
1. Have an operational computer of server quality.
2. Have an addressing system that is independent of the DNS.
3. Able to cope with variable connectivity.

These features allow sharing of computing resources via direct exchange between computers.

In general it is agreed that there are two major forms for this new paradigm:

*HYBRID: Some nodes are Router-terminals that facilitate the interconnections between Peers.

*PURE: All nodes are Peers, and each Peer may function as router, client, or server, according to the status of the query.

SELF-ORGANIZATION: Peer-to-Peer architecture generates its own organization for its nodes. Some of the suggested impacts of P2P communications are related to self-organization. This may occur on two levels:

1. With respect to a given P2P community, any computer granted permission to become a node in the community is, at the same time, granted equality to every other node in the community. Adding peers to a network requires no re-organization, central or otherwise.

2. On a larger scale, say billions, the P2P community can naturally organize (according to interests) into millions of smaller SIG virtual webs. In effect, peer node organization is independent of whether an individual node is connected or not. This is the feature that handles variable connectivity. This is the same feature that facilitates scalability.

This self-organizing facility is an important difference from the centralized organization of client/server computing.

CONGESTION: Because the heavy transfers are at the users edge of the network and Peer-to-Peer, congestion is minimized.

SCALABILITY: With lowered congestion and automatic organization, scalability becomes less complicated and highly facilitated by the architecture.

--Lecture Notes by Ross Lee Graham

P2P Home

Ross Lee Graham
Last modified: Thu Apr 19 19:39:16 MET DST 2001