There are many tools and techniques to avoid unwanted e-mail – the kind of messages pushing products like pornography, prescriptions and printer cartridges.
But a popular technology called peer-to-peer networking – which made Napster famous – is providing a new way for computer users to slice through spam.
Cloudmark, a company in San Francisco, is giving e-mail users the ability to vote on what is spam and then update spam-blocking filters across a peer-to-peer network called SpamNet, which is similar to the systems people use to trade digital music over the Internet.
SpamNet draws on some of the strengths – philosophical and technological – that made Napster such a powerful program and applies them to a cause that everyone can rally around. As the network’s spam-fighting community recruits more members, the database will grow and sustain a living filter that evolves to intercept spam.
“It’s allowing users to take control of their spam situation,” said Karl Jacob, chief executive officer of Cloudmark. “By definition it will get better over time.”
SpamNet – which is the creation of Napster co-founder Jordan Ritter and programmer Vipul Ved Prakash – links users to the network through a free software plug-in for Microsoft’s Outlook e-mail software. SpamNet automatically checks incoming e-mail against its database of known spam messages and attempts to filter out the offending material.
All e-mail is still downloaded to the computer, but messages identified as spam are sent to a “spam” folder for perusing or deleting.
The program – which is available for Outlook 2000, 2002 and XP – installs “block” and “unblock” buttons within the e-mail software. If an unwanted message slips through, the user highlights it and clicks the block button. If enough users flag the same piece of e-mail as spam, SpamNet will stop it from reaching everyone’s e-mail in-box.
If SpamNet stops a legitimate piece of e-mail, users can find it in their spam folder and click the unblock button, which moves the message – and future similar e-mail – to the in-box.
If other SpamNet members do the same, those messages will flow again to all e-mail in-boxes within the peer-to-peer network.
To protect against spammers undermining the system, SpamNet has a “truth evaluation system” that is weighted toward “trusted” members who consistently block spam that most others are also blocking. By giving priority to those users, the system keeps spamming spies from trying to unblock mail.
“It’s hard to gain trust in the system and very easy to lose it,” Jacob said. “Your reputation goes down very, very quickly, and the system is very effective at rooting out people trying to work against it.”
SpamNet focuses on the content of the unwanted e-mail messages instead of who sent them or where they came from.
The system calculates a fingerprint from the content of the message and assigns every e-mail a unique identifying number, and those numbers are compared with SpamNet’s database to protect privacy. Because these codes are a small amount of data, the SpamNet comparisons can be done quickly, Jacob said.
“Spammers have gotten really smart” in covering their tracks, Jacob said. “But the body of the message is what the spammers care about. Oddly enough, that is what they can’t change too much without losing the content of the message.”
Jacob said tests of SpamNet, which went online in late June, show users can reduce their daily flow of unwanted mail by more than 90 percent. The program has been downloaded more than 87,000 times, and there are nearly 40,000 users helping the system identify spam.SpamNet members praise the program, but some have complained of flaws in the peer-to-peer networking approach.
Some SpamNet members who had signed up for bulk mailings from Web sites or companies are now blocking the messages, causing problems for other SpamNet members who want the mailings.
According to some SpamNet members, bulk e-mailings from companies like Microsoft and Lucent Technologies have been branded as spam and sent unceremoniously to the spam folder.