Security startup wages continuous war games against networks

Startup SafeBreach automatically assesses corporate networks to find out whether they offer up enough security loopholes for real-world attacks to succeed.

Using software probes called simulators distributed throughout customers’ networks, SafeBreach attempts to establish connections among devices and network segments just as a hacker would do in trying to carry out malicious activity.

These automated attempts are driven by the Hacker’s Playbook, a SafeBreach library of known attack methods that the simulators try in order to discover weaknesses and reveal how these vulnerabilities might be exploited to carry out successful breaches.

So simulators might find individual weaknesses in a desktop Internet connection, a credit card database and a management platform that could be strung together to nab customer credit card data. This would be reported on a single screen.

In this way corporate security pros can discover and close previously unrecognized security gaps and also validate that the controls they have put in place actually do what they were designed to do or whether they are rendered vulnerable by misconfiguration.

The tests run by the simulators are orchestrated centrally, driven by algorithms that use the results of initial probes to determine what methods to try next, says Guy Bejerano, the CEO and a co-founder of SafeBreach. “It’s just like a hacker would try,” he says.

SafeBreach differs from penetration testing in that it offers ongoing tests of networks as they change over time vs. periodic snapshots that leave long time gaps for potential vulnerabilities to be exploited. Constant software updates and device reconfigurations as well as network architecture changes mean potential attack paths are in constant flux, Bejerano says.

MORE:10 Young Security Companies to Watch

The platform can be used to confirm that networks meet security standards set by regulators and industry standards and ensure that networks are securely segmented to prevent access that could lead to widespread breaches.

Dedicated security staff could approximate what SafeBreach does, but qualified personnel is hard to find. “Hiring hackers is cool but it’s not scalable,” Bejerano says. “It would take too long and is super expensive.”

He says the company developed its Hacker’s Playbook using the knowledge of its researchers, some of them former Israeli military, who compile known and lesser known ways to access information and toss in ideas of their own. They are led by CTO and co-founder Itzik Kotler. The company plans to set up a community SDK that would let customers contribute to the playbook, Bejerano says.

The technology has been named one of 10 finalists in the Innovation Sandbox contest at the RSA Conference 2016, which seeks the most innovative security startup.

SafeBreach is scheduled to be available in the fourth quarter. Pricing is based on the number of simulators deployed. It costs $50,000 per year for 10 of them, and that includes a service-level agreement. The platform can be deployed either as software as a service or on premises.

The company is based in Menlo Park, Calif., and is backed by Sequoia Capital and Shlomo Kramer, a co-founder of Check Point, Imperva and Cato Networks.