Is it all happening again? Nervous computer executives – many of whom missed the significance of the Internet browser – spent last year worrying about peer-to-peer (a term actually around for years basically meaning connecting up resources, typically two computer servers, where both act as equal partners with similar capabilities). Was it just a somewhat chaotic and slightly subversive playground for fans of chat rooms and music piracy, or something of importance for the industry at large?

What prompted their interest was the rise of two distinct applications that would (after the event) both be labeled peer-to-peer: Instant Messaging (IM) and Napster. IM first appeared as a free Internet chat program called ICQ in November 1996, developed by an Israeli fir, Mirabilis. Barely a year later, ICQ had over seven million subscribers, some two million of whom were using it daily.

That caught the attention of AOL, which had already developed similar technology for its own proprietary network. When it saw how fast ICQ was growing, AOL promptly adapted its own technology to work over the Internet and, in June 1998, actually acquired ICQ for $287 million. By then there were some 10 million ICQ users. Today, AOL has something approaching 30 million users in the US for its combined IM services, and Microsoft and Yahoo have a further 20 million IM users.

The rise of P2P
Napster was developed in 1999, and very soon colleges were finding that their networks were being overwhelmed with students swapping MP3 music files. By the end of last year, Napster claimed to have 32 million users, with almost one million of them connected at any one time.

That’s an impressive acceptance rate, and it can’t be ignored. Both IM and Napster are currently free programs with very specific and limited functionality. Can the peer-to-peer model be extended to other application areas? And does it make any sense to do so?

Peer-to-peer networks are ‘transient,’ in that they don’t expect a device to have a permanent presence on the network. That’s why both IM and Napster bypass the Internet’s established Domain Name System (DNS) in favor of their own address schemes. DNS assumes a permanent IP (Internet protocol) address, whereas most home PCs and mobile devices dial-in through modems, and use a temporary or dynamic IP address, which is different every time they log in. In effect, the long-established DNS system had been freezing out the huge numbers of transient users from full participation in the Web.

Clay Sharkey, a partner at technology analyst firm, Accelerator Group, defines the peer-to-peer label as “a class of applications that takes advantage of resources – storage, cycles, content, human presence – available at the edges of the Internet.” As he points out, hundreds of millions of devices – not just PCs, but increasingly phones, personal digital assistants and an emerging set of Internet ‘appliances’ – have been connected up in the last few years, representing a huge pool of largely unused resources. Such devices have typically acted as one-way ‘clients’ for passively viewing material on the Web. Peer-to-peer technology has the potential to turn many of these into ‘servers’ as well as clients, contributing resource and information to the network, as well as consuming it.

Meanwhile, the development of e-commerce and business exchanges show that the Internet is gradually evolving from the simple presentation of information to a platform for running applications in its own right – or as others have put it, the “two-way Web.” Two technologies, XML and SOAP (simple object access protocol) – are helping to make this happen. The XML markup language opens up direct access to structured data within a Web page, making it programmable in a way somewhat akin to a database. And SOAP provides the means for XML to travel over the Internet. Microsoft, IBM and Sun are all supporting both technologies.

Companies to watch
Now that momentum has been built, the commercial startups are emerging with heavyweight applications. One particularly promising sector is financial services. Rob Hegarty, of IT research consultancy TowerGroup, expects to see “widespread adoption of P2P technologies in all facets of financial services, particularly in investment research and electronic marketplaces.” He points to two firms, WorldStreet and LiquidNet, which are already entering the investment management market.

Others to look out for include collaboration software house Groove Networks, storage networking companies such as FilePool and MangoSoft, and distributed systems infrastructure firms such as Centrata. Microsoft, Intel, Exodus and other large players are keeping a close watch and working on some applications of their own.

The future of P2P
But there are hurdles to overcome. Pioneers must do battle with security worries, reliability, systems management difficulties and potential performance bottlenecks that users of free services such as Napster and IM have had no choice but to accept. Paying customers are likely to demand much more.
It’s hard to size the peer-to-peer market, because it can potentially cover so many different areas. Shirkey estimates that, taking conservative estimates, the world’s Net-connected PCs currently host “an aggregate ten billion MHz of processing power and ten thousand terabytes of storage, assuming only 100 million PCs among the net’s 300 million users, and only a 100 MHz chip and 100 MB drive on the average PC.”

But as Shirkey himself points out, users won’t decentralize for decentralization’s sake. They will adopt only enough of the technology to create new functionality, or to improve their existing set-ups.