In the Year 2005, Will Your Anti-Spam Arsenal Be the Same? - Page 2

 By Jacqueline Emigh
Page 2 of 2   |  Back to Page 1
Print Article

Bigger Plans Ahead

Regardless of their current stature, many anti-spam vendors are mapping larger plans. MailFrontier, for example, brought out an anti-spam gateway earlier this year after taking its first step into anti-spam in 2002 with a desktop plug-in for Outlook.

MailFrontier's Matador client for Outlook added support for Outlook Express, Hotmail, and IMAP in February. As differentiators for Matador, MailFrontier points to collaborative filtering -- based on a conglomeration of user-identified spam -- and "e-mail challenges," meant to distinguish machine- from human-generated mail. "With the e-mail challenges, the sender might receive a picture back, and be asked to identify, 'How many kittens?'" illustrates Grey.

Further down the road, MailFrontier envisions adding other capabilities for mail management, including policy management, lifecycle administration, and "trusted network and other methods of certifying inter-company and company-consumer communication," according to Pavni Diwanji, company CEO and co-founder.

For its part, anti-spam vendor and service provider BrightMail lures spammers to specially set up e-mail addresses called "honeypots." Another free tool, Jackpot, is an SMTP honeypot application for Windows.

As mentioned before, many commercial features either have roots or parallels in freeware products, yet many administrators prefer the relative simplicity of shrinkwrapped products or subscription services (or both). "We purchased SurfControl E-Mail Filter due to the overwhelming amount of spam clogging our network. The product works extremely well and helps protect our company from junk e-mail everyday," claims Albert Rodriguez, IT Manager at ImageMaster Financial Publishing, Inc.

Illinois Tool Works, on the other hand, has installed two of Mirapoint's MessageDirector anti-spam hardware devices, reports Marc Pilano, IT director for the Fortune 200 manufacturing firm. Pilano thinks the devices, which sit at the edge of the network, will help prevent integration and licensing issues that might come up if each company division deployed its own anti-spam solutions.

Beyond the free tools, vendors new to the anti-spam scene face considerable competition already. Aside from those companies already mentioned, administrators can choose commercial anti-spam hardware or software gateway solutions from BorderWare Technologies, Postini, CipherTrust, Clearswift, Elron Software, JunkJam, Tumbleweed Communications, and Sendmail, with its Mailstream Manager.

In the crowded anti-spam client space, Mail-Filters, Sunbelt, and MailWasher are a few of the companies not previously mentioned. Many are offering 30-day free trials.

Technologies Will Broaden, but the Field Will Narrow

Analysts agree, though, that the anti-spam tools available to customers will change dramatically over the next few years.

"By 2005, there will be no more than five leading providers of anti-spam logic. Winning in the anti-spam game requires a focused and specialized lab, not unlike an antivirus lab, where observations, algorithms, and heuristics are applied to identify spam messages," says Gartner's Grey.

"Vendors that become complacent will lose. Just as Tumbleweed, Sendmail, and many others welcome signature subscriptions from your choice of antivirus lab, so in the future such products will offer a subscription to anti-spam logic from one of the leading anti-spam logic vendors."

Dan Keldsen, an analyst at Summit Strategies, predicts that some of the current anti-spam vendors will spin out in the direction of total e-mail management, adding features such as e-mail encryption.

Bayesian Filters and Machine Learning

Meanwhile, other new technologies will bubble forth from anti-spam research labs. Many products available today rely on wellworn standbys like black lists and keyword comparisons.

Newer products are adding heuristic (pattern-matching) analysis as well. One intriguing area of research right now revolves around Bayesian filters, which assign spam probabilities to individual words in e-mail messages. For example, the filters are capable of recognizing "words" that use a combination of numbers and letters -- or a string of nonsensical punctuation marks -- to evade detection by other means.

Researchers are also exploring various machine-learning technologies. At e-lab Bouygues SA in France, developers are creating Smartlook, an assistant for Microsoft Outlook aimed at learning users' mail filing habits in order to detect and automatically filter out unwanted messages.

Whether you prefer free solutions or commercialware, you should keep your eyes peeled on what developers are doing. The anti-spam landscape keeps changing so fast that you never know what will crop up next.

» See All Articles by Columnist Jacqueline Emigh

This article was originally published on Mar 24, 2003
Get the Latest Scoop with Networking Update Newsletter