Everything Old Is New Again

In computing's mainstream, middleware, along with client/server computing, has always been fundamentally about small boxes working vertically with bigger boxes, or servers.

In computings mainstream, middleware, along with client/server computing, has always been fundamentally about small boxes working vertically with bigger boxes, or servers. Historically, there was always another choice: peer-to-peer (P2P) computing.

P2P, the mainstay of such once-popular low-end networks as Artisofts LANtastic and Novells Personal NetWare, fell into disuse, as the greater speed, security and manageability of server operating systems took hold.

Of course, P2P never fell completely out of favor. P2P networking comes with Windows 9x. P2P also is the engine behind well-known videoconferencing and real-time interaction programs like CUseeMe and Microsofts NetMeeting.

Today, its back to the future for P2P, as popular distributed file-sharing applications such as Napster and Gnutella show that new-fangled, but old-fashioned, P2P file sharing still can be compelling. The new part is that with modern P2P, there is no file or resource directories as such; instead, resources are dispersed into serverless databases that can be accessed with P2P applications.

These next-generation P2P applications rely upon the Internet staples of Web servers and single-purpose specialized browser applications. Data and files tend to be kept in simple distributed databases with a minimal (Napster) central database for coordination or, in the case of Gnutella, no centralized information base whatsoever.

The new P2P has several advantages. The least relevant for business is user anonymity. MP3 song traders love it, but it has little business use. However, the other advantages—transaction speed, no single point of failure and improved network performance—could convince business users to give P2P a try. Another point in P2Ps favor is that it doesnt require server capital expenses. For most P2P "servers," well-aged Pentiums do just fine.

The business question is whether P2P will become compelling again for our customers business uses. Some observers, like Anne Zieger, chief analyst for Peer to Peer Central (www.peertopeercentral.com), dont believe that simply revitalizing the old P2P file concept will work for business.

Instead, Zieger, borrowing from Roku Technologies definition, describes modern P2P as "a networking architecture that facilitates direct person-to-person communication." Intels more conventional definition of P2P as "sharing of computer resources and services by direct exchange," she believes, is "too general a description to benefit technology buyers."

Instead, Zieger sees P2P business opportunities in four functional areas: "managing and sharing information, collaboration, networking/infrastructure and e-commerce enablement."

The first is good old P2P file sharing. Collaboration seeks to enable groupware without the expense of centralized groupware programs like Lotus Domino. Groove Networks Groove is the unquestioned leader in this arena. Some of the network infrastructure crew, like myCIO.com, are looking at P2P to enable software updating. Others, such as Cytaq, are planning to use P2P to enable scalable, cross-platform resource sharing across the enterprise.

E-commerce, though, might prove the most profitable P2P area. In particular, B2B players that have grown disenchanted with server-based B2B exchanges are beginning to look at P2P exchange alternatives. Companies like Biz2Peer are exploring P2P to enable B2B and supply-chain transactions. Others, such as Syndicator from AgentWare, already are being used in business. The thought is that P2P can provide faster one-on-one transactions with less equipment.