Peer-to-peer technology may have pulled off the greatest disappearing act of the post-dot-com era. Once heralded as the second coming of the Internet, file-sharing giant Napster was sued out of business, and many of its descendants, such as Kazaa and Morpheus, are under fire from the Recording Industry Association of America (RIAA).
Grizzled veterans of P2P technology, such as Andy Oram, editor of “Peer-to-Peer: Harnessing the Power of Disruptive Technologies,” shake their heads in disappointment. “Honestly, I’ve felt for a while now like it was part of my past,” Oram says of the technology he once passionately embraced.
Nonetheless, the innovations unleashed by Napster are making modest headway in corporate networks. For example, network administrators are discovering that distributing some file-sharing capabilities out to the edges of a network can make their workforce more versatile and mobile. Sun Microsystems has launched a set of P2P tools, as has IBM, though they have yet to catch fire in the marketplace.
Peer-based technologies may yet change the way corporate computing works. Just remember: Don’t call it “P2P” when you are selling it to executives. That term is tainted.
What Is It, Exactly?
Evaluating P2P in the enterprise is an ambiguous endeavor. What does the term really encompass? For some, the image of file-sharing programs like BearShare, Limewire and Napster is most prominent. For others, it is about “presence” — the ability of a computer user to sense when a coworker may be on the network — though that feature is more properly associated with instant messaging programs than with P2P. Still others might say P2P should promote distributed processing power within the enterprise, just as the [email protected] project uses volunteers’ Pentiums to search for extraterrestrial life.
In fact, it turns out that the features commonly deemed most important about P2P — decentralized file exchanges, non-hierarchical communications, direct collaboration and distributed processing power — are rather unimportant to most corporate buyers. This dilemma is familiar to Ray Ozzie, formerly the brains behind the Lotus Notes program, who now runs P2P software venture Groove Networks. Lotus Notes offered sophisticated file sharing in an age before the World Wide Web, but most corporate customers ended up using it for little more than e-mail. Likewise, Groove is finding that the killer P2P app, collaboration, is not so killer.
“There’s about 3 million lines of code in Groove, but it’s the last 250,000 lines, the part that you use to develop applications internally, that matters to customers,” Andrew Mahon, director of strategic marketing at Groove Networks, told the E-Commerce Times.
Groove Workspace, a peer-based file-sharing program, was a proof of concept built on top of the company’s P2P toolkit, he said. Businesses can start using it for US$69 per seat, putting it within reach of even small companies.However, three pieces of application infrastructure have made the underlying Groove code libraries more important to users than Workspace itself. They are: synchronization of changes made to documents offline; the ability to apply strong encryption to files with sensitive content; and the ability for remote workers to connect to the enterprise through virtual private networks.
In the Groove
One prominent Groove customer said those three factors are exactly why he implemented the software. Craig Samuel is chief knowledge officer in charge of information sharing among the 65,000 employees in Hewlett-Packard’s services division, which is spread across 160 countries. “The business problem we try to solve internally is, how do you get large organizations around the world to access and share the right knowledge in a very responsive way?” Samuel told the E-Commerce Times.
For HP, Samuel said, achieving this goal meant integrating peer-to-peer technology with the Web-based Sharepoint package used widely throughout the company, which in turn required using Groove’s capabilities as modules to be plugged into Sharepoint. “We wanted to be able to not just share information around the periphery of the organization,” as in Napster-style file sharing, “but also to connect that to centralized programs.” Specifically, Groove offered a better way to synchronize data than Microsoft Outlook, according to Samuel, and it was an easy way to add 192-bit encryption to file sharing.
As a customer since version 1.0, HP’s experience helped shape Groove. “In the beginning, Groove was just a nice collaboration tool,” Samuel said. “Now, people can work together in a common space, in portals, and push content between different repositories. That’s been the key” to making Groove a productivity enhancer. In the future, he added, HP may build more complex applications on top of Groove using Microsoft’s Visual Studio .NET.
However, although Groove is cherished by some customers for offline synchronization, its popularity only highlights the shortcomings that hobble P2P in general. P2P guru Oram said file-sharing programs and the like have “serious problems.”
“There are three things that P2P really lacks, namely a consistent addressing scheme, metadata and a way to ensure the reputation of participants in the network,” he told the E-Commerce Times. Because of those shortcomings, “right now, it’s hard to say P2P makes things better for the end user; it makes things more difficult.”
On the first point, Oram noted that although the presence features of Jabber, an instant messaging program, are good, he would like to see a consistent protocol for finding and addressing P2P participants, akin to plain old e-mail addresses. He added that he thinks metadata is needed to replace the unstructured, stream-of-consciousness nature of instant messaging with an organized method of exchanging complex documents. Lastly, Oram argued that as P2P technology moves beyond stealing music, there must be a way to validate the reputation and trustworthiness of users participating in the network.
Others agree that beneath all the fluff about P2P, today’s programs are immature. “There are mixed emotions about this in IT,” said Bob Parker, an analyst at AMR Research. He told the E-Commerce Times that although peer-based software can be an effective tool for ad-hoc collaboration, companies are worried that they will not be able to properly archive materials traded back and forth among peers.
“Any communication in a bank related to an equity that might be traded comes under purview of [the] SEC, and that needs to be archived,” Parker said. Of course, third-party tools can be added, but “out of the box, P2P has these archival issues.” He agreed with Oram about the issue of addressing, to say nothing of the headaches created on networks flooded with decentralized file sharing.
Is It Middleware?
However, if companies like Groove can succeed despite the pressing issues surrounding peer-to-peer technology, one can only conclude that P2P has not been rejected by the market. Far from it. The technology appears to be finding a niche by addressing existing communication issues in enterprise computing. Such unglamorous hacks used to be called “middleware,” a kind of software piping sold by established companies like IBM.
However, Groove’s Mahon eschews that label, not least because it threatens to pull Groove into Microsoft’s rapacious maw as the software giant seeks to incorporate more middleware into Windows. (Microsoft has an investment in Groove.) Instead, Mahon is stumping for the liberating power of the technology. He believes that, like Napster, Groove brings a little bit of an anarchic quality to the way an increasingly mobile and connected staff functions inside the Fortune 500.
“Users want to just work together without having to have permission, just as they do with e-mail,” he said. “Groove lets IT maintain control over policy, while users can control the content that gets exchanged.” That may be a good enough selling point for executives — so long as they are not spooked by visions of illicit file swapping.