Using a given user's network behavior, the Dtex platform businesses can figure out actions such as whether the employee is thinking about leaving the business. The system then could alert administrators to take steps to make sure the employee doesn't walk away with sensitive information.
+[Also on Network World: Beware the insider threat; Anthem's latest breach estimate says 78.8 million were affected]+
At the same time the system works without gathering personal data from the endpoints that would violate privacy laws, even the strictest ones in Germany, he says.
The company distinguishes itself from competitors in that it has a lightweight client that grabs endpoint data in increments to report changes as opposed to sending a full scan of the endpoint each time, making more efficient use of the device's processing power, say Avivah Litan, a vice president at Gartner.
The company falls into the category Gartner calls User Behavior Analytics. It competes against the likes of SpectorSoft, ObserveIT and Rapid 7, she says. Some competitors have no endpoint clients at all but gather fewer details. The marked for this type of insider threat detection is very immature, she says, but the products are in high demand, especially among the largest corporations that seek more analytics around their security.
Dtex started up in Australia and grew gradually into markets in Asia and Europe, and is announcing a move to San Jose, Calif., where it this week got an infusion of $15 million in venture capital to expand its U.S. presence. Investors are Norwest Venture Partners and Wing Ventures.
The company will use the investment to hire sales people and beef up the engineering team, Koo says. The company has been profitable in Asia and Europe and expects to fall out of profitability this year because of the expansion, but that it will rapidly regain a positive bottom line.
Its technology has three parts: a micro-agent deployed on endpoints, a server that gathers the data from the endpoints and an analytics engine that sorts it all out.
The client is lean due to work the company did for early customers who had low-powered endpoint machines and 56K bps dialup links to some of their branches. Dtex pared down the software to gather only essential information and use less network bandwidth.
The system monitors users' behavior for 90 days to establish baseline behavior then seeks aberrations from it. "It's easy to see people who are leaving with market-sensitive data," he says. The company has created 330 behavior profile rule sets based on actual incidents that flag suspicious behavior, and that helps the system work with low false positives. That is more difficult when it runs across an incident it has never seen before, he says.
If a sensitive incident arises that involves law enforcement, customers generally prefer to have Dtex techs investigate it as a third party. Many customers who have run the server themselves have opted to move to a managed service after seeing how rapidly Dtex can run an investigation, Koo says.
The company is identifying itself as detecting insider threats as opposed to finding malware that's running exploits from endpoint devices, Koo says. Other companies specialize in that, although more and more Dtex is looking at that area.