Proofpoint improves spam prevention

Upgraded MLX technology teaches software to detect more junk mail.

Anti-spam start-up Proofpoint last week released an upgrade to its corporate gateway software that promises to give customers new ways to detect and deflect spam.

Anti-spam start-up Proofpoint last week released an upgrade to its corporate gateway software that promises to give customers new ways to detect and deflect spam.

Proofpoint's Protection Server features the company's MLX technology, a machine- learning system that "teaches" the software to identify spam based on several attributes. These attributes include telltale signs in an unwanted e-mail's IP address, message headers, structure, content and phrasing. For example, spammers often embed an abundance of HTML code in a message's text - Proofpoint's Protection Server automatically will detect its presence and flag the message as potential spam.

MLX used 10,000 attributes in Version 1.0 of Proofpoint Protection Server; with Version 1.5 there are more than 50,000 attributes used to flag spam, says Greg Smirin, Proofpoint's vice president of marketing and products.

"It makes a big difference when you have a whole lot more of these attributes, because you can take a much more fine-grained approach to distinguishing [legitimate] messages from spam," Smirin says.

Tire reseller TBC has experienced a noticeable drop in spam after running Proofpoint Protection Server 1.5 for just a few days, says Mark Beal, senior director of IT. "Before the upgrade I was getting anywhere from eight to 10 spam messages a day," says Beal, who upgraded from Version 1.2. "I haven't gotten any over the last four days."

TBC made the decision to upgrade based on the previous version's performance, he says. "After seeing how well the product was working and some of the enhancements in 1.5, I didn't even think twice about upgrading," Beal says. "The biggest thing [about the upgrade] is the additional spam blocking. Before some was still getting through the filter for whatever reason, but now it's virtually eliminated."

The machine-learning algorithms Proofpoint Protection Server uses to catch spam are similar to the techniques used in genomic sequencing analysis, says Andres Kohn, Proofpoint's director of product marketing. The software analyzes millions of e-mails and, based on the attributes found in spam messages caught, generates models of what spam messages look like. This way the software can catch unwanted messages flowing into corporate e-mail systems without human intervention, Kohn says.

Other enhancements in Version 1.5 include security-audit compliance, so that the software works within the parameters of a company's security policies, and new reporting options that can be customized and integrated with other reporting tools.

Proofpoint Protection Server 1.5 costs between $2 and $20 per user annually, depending upon the number of people in an organization.

Proofpoint released Version 1.0 of its Protection Server in July. The company competes with anti-spam gateway software vendors including Brightmail and Cloudmark, as well as with security vendors that have added anti-spam modules to their offerings such as Trend Micro and Network Associates. Eric Hahn, former CTO of Netscape Communications, founded Proofpoint.

Editors' Picks
Join the discussion
Be the first to comment on this article. Our Commenting Policies