Scared stiff by daily reports of new and ever-more-sophisticated attacks on 802.11 wireless LAN networks, enterprise IT managers are turning to the new breed of wireless intrusion detection systems (IDS) that are fast becoming a must-have feature of any self-respecting wireless LAN vendor's portfolio.

Wireless intrusion detection systems work by continuously scanning an enterprise's airspace for the tell-tale signatures that indicate sophisticated denial of service and man-in-the-middle attacks against networks secured by 802.1X-based authentication mechanisms and/or VPN tunnels are underway.

As such, wireless IDS is a step beyond the rogue detection capabilities that are now promoted by many vendors of business-grade equipment. However, recent IDS product announcements from the likes of Aruba Wireless Networks, IBM Corp. (NYSE: IBM), and Red-M suggest that there's more than one way to skin a hacker.

Aruba, for example, offers IDS as a software application that runs on its core or edge wireless LAN switches (see Aruba's Mini-Switch), while IBM offers a managed wireless IDS service for enterprises that don’t want the hassle of doing it themselves (see IBM Eyes the Enterprise). Red-M, meanwhile, says IDS can either be run in-house or be outsourced to a third party (see Red-M Launches IDS).

According to Red-M's president and CEO, Karl Feilder, there are basically three ways to do wireless IDS. The first is to use the existing access point network to monitor traffic patterns. This is the approach being taken by switch startup Airespace Inc., which has seven or eight as-yet-unannounced customers using its system for IDS, according to Alan Cohen, its VP of marketing.

The second way is to dedicate passive 802.11 monitors to tracking all wireless traffic and then analyze the data at a central server for attack patterns. This is the approach being taken by AirMagnet Inc., Aruba, IBM, and others.

Red-M favors a third way: "We have our own specially-designed sensors that analyze traffic, identify the fingerprints of the bad guys, and then send anything suspicious back to the server," says Feilder. He claims that this approach is "more scaleable" because most of the processing is "done at the edge of the network."

Despite all this, Meta Group analyst Chris Kozup says customers shopping for IDS should really look for systems that allow manual control and adjustment of attack patterns and the resultant security reactions, rather than autonomous systems that put security lockdowns in place without user intervention. "Administrators don't want automatic systems," Kozup contends.

Red-M's Feilder cites banks, insurance companies, healthcare agencies, and car manufacturers among the customers using Red-M’s wireless IDS. Ironically, however, he says: "We’re selling most of our products to companies that want to ensure a no-wireless policy." Spoilsports.

— Gabriel Brown, Research Analyst, Unstrung