Cloudmark Unveils P2P Anti-Spam System

The co-founder of Napster, Jordan Ritter, is putting P2P technology to work fighting the growing scourge afflicting Internet users and hindering legitimate email marketing — spam.

“Imagine taking the power of the tens of millions of people that formed a community around the Napster network and using that network effect to fight back against spam,” said Ritter, who has founded the spam-fighting San Francisco-based start-up called Cloudmark, where he serves as chief technology officer.

That community — which the company has dubbed SpamNet — is being built by way of a free add-on to Microsoft Outlook 2000 or XP, a beta version of which the company is distributing from its site. The system works by asking members to report any spam messages that they receive, which they can do via buttons that appear in Outlook when the add-on is installed. A report then makes its way back to a server running a program called Vipul’s Razor. That server then passes the word to other servers running the program. The result is that end users are protected from receiving any spam messages that have been reported by other network members. The messages are filtered into a spam folder on Outlook, so people can ensure that email isn’t being filtered incorrectly.

Similarly, users can report legitimate messages that mistakenly end up in their spam folders, using the same buttons that appear in Outlook. The company says this feature helps legitimate marketers distinguish themselves from spammers, because users can indicate that they opted-in for something that was erroneously filtered into their spam folder.

“They are able to say ‘I like this,’ basically, or not vote that it’s spam. That creates a situation in which a majority of people say it’s not spam; so it’s not spam,” said Karl Jacob, chief executive officer of Cloudmark. For the last two years, Cloudmark has operated a network of Unix users via Razor, but this Outlook add-on, because it requires less technical knowledge from the end user, exponentially increases the number of people that could be involved.

Another safeguard against false reports is Cloudmark’s “Truth Evaluation System,” which follows up on spam reports and begins to assign individuals with a reliability score, based on their propensity to issue true — or false — spam reports. That reliability score is taken into consideration when individuals make future reports.

“It works a lot like eBay’s reputation system or rating system,” said Jacob. “It watches people over time and sees whether other people agree that what’s reported is actually spam.”

The system is similar to that employed by popular anti-spam solution SpamCop, in that a “blacklist” is established using users’ reports. The difference, however, is that the SpamNet system is theoretically more scaleable because of its P2P architecture. Additionally, the SpamNet “blacklist” consists of individual messages, rather than objectionable IP addresses — which SpamCop blocks if it sees a certain percentage of spam reports relative to the volume of mail that’s sent via that server.

Cloudmark apparently hopes to monetize its solution by charging ISPs and corporations to use its spam-blocking filters, while allowing most individuals — spam reporters — to use the service free of charge. It says it intends to release a version for Outlook Express, as well as for other email clients. A for-pay version of the service is also in the works.

The company is launching into an increasingly crowded space, occupied by firms such as Brightmail, McAfee, MessageLabs, and SpamCop, among many others. As the volume of spam has increased — Brightmail says it has expanded six fold in the last six months — corporations, individual users, and legitimate email marketers have grown increasingly concerned about the issue.