Peer-to-peer networking has caused a dramatic increase in Internet traffic. Today, the major P2P user is the consumer, and the major application is media file sharing. Music, films and the like now are available to consumers using P2P technology. Depending on the analysis methodology, P2P networking accounts for 60 percent to 89 percent of all Internet traffic.
P2P uses the computing power at the edge of a connection rather than within the network. The concept of clients or servers does not exist. Instead, peer nodes function as both clients and servers to other network nodes.
Although pure P2P networks exist, almost all efficient P2P networks use a hybrid approach in which the critical applications of indexing and searching are implemented in a client/server form, and data transfer is accomplished in a P2P manner. This technology is analogous to a router's use of a route or table look-up processor and multiple forwarding processors.
There is a fundamental difference between P2P and client/server network traffic. For example, when using the classic client/server Internet FTP application, a user uploads a file to the FTP server and then other users can download that file, without any user-to-user communication. The only network bandwidth being consumed is that of the client and the contended server. As the number of clients increases, the available server bandwidth decreases. If the same file distribution application is implemented using P2P networking, the download bandwidth increases with the number of distributed peer nodes.
Don't think it's just for consumers
Most corporate networkers believe P2P, with its ability to hog bandwidth and its myriad security issues, is a consumer phenomenon. This is a misconception that in the future may be the undoing of the corporate network. P2P network applications such as KaZaA or Napster may not have a place in the corporate environment, but the same cannot be said for BitTorrent's File Sharing technology, currently being used for Linux software distribution, and Groove Networks' Virtual Office application, designed for shared workspace/online collaboration.
The use of BitTorrent and Groove software currently does not impose a significant bandwidth demand on corporate networks. But Microsoft's recent acquisition of Groove may drastically change the bandwidth consumption equation. One of Microsoft's stated intentions is to add Groove's P2P technology to its next-generation operating system called Longhorn. With Longhorn's arrival on the desktop and server, Microsoft may single-handedly redistribute and accelerate corporate bandwidth demand.
Grid computing is P2P too
Also on the horizon is another corporate networking nightmare called the grid. Lo and behold, the underpinning of all grid technology is P2P. With the adoption of Globus 4.0 as the new XML-based protocol standard, grid services will become the P2P of Web services.
As corporate terminal-mainframe centralised networking evolved into client/server distributed networking, so shall all forms of corporate networking eventually evolve into P2P. The networking traffic dynamics are not trivial in the P2P world. Complexity is the norm, with today's clients and servers becoming next-generation IT peer nodes.
This seems to be a natural blueprint for the future. All edge nodes will become highly intelligent, simultaneously and randomly capable of dynamically participating in the combined execution of an application as well as interfacing directly with the user. The corporate network also will become more internally intelligent, being required to optimise applications bandwidth demand dynamically while simultaneously managing network latency/jitter and corporate policy.
Be forewarned, get educated and be prepared for the network implications of corporate IT P2P applications. The corporate next-generation network future may be just around the corner in 2006.
Dzubeck is president of Communications Network Architects, an industry analysis firm in Washington DC. He can be reached at [email protected]