Attack mitigation tools fall short, security vendors say

When preparing a security plan, enterprises should assume they will be breached at some point

SAN FRANCISCO -- Acknowledging that security technologies to prevent cyberattacks aren't always up to the task, several vendors at the RSA Conference here advised companies that are making security plans to just assume that they will be breached at some point.

Rather than pouring resources into stopping all attacks, the better strategy is to acknowledge that some attacks will inevitably penetrate their defenses, they said. Therefore, the goal of any enterprise security strategy is not to focus solely on attack mitigation, but also on quick detection and response.

"The typical focus today is on trying to prevent malware from getting in through the front door," said Bret Hartman, chief technology officer at RSA, the security division of EMC. "The problem with that approach is that there's always a percentage [of malware] that does make it through. There's been an overemphasis on infiltration. The goal is to shift focus and assume that you have been infiltrated."

Such advice signals an epiphany of sorts in an industry where vendors have always insisted that their technologies, if properly deployed, would protect companies from attacks. Events over the past year, such as the attacks on Google and those tied to Stuxnet, have highlighted the fact that it may be impossible to fend off every determined adversary.

The vast and ever-increasing amounts of data that companies need to manage and the innumerable ways in which that data can be accessed have greatly heightened the need for companies to look beyond traditional defense strategies. While such defenses are useful in blocking about 75% of the threats out there, new approaches are required for dealing with the remaining threats, experts said.

Exacerbating the problem is the increasing sophistication of attack weapons and approaches, vendors said. Many of the malware tools that companies need to deal with these days have been explicitly designed to evade detection and to remain hidden for long periods. Once such tools infiltrate a network, they are almost impossible to detect and eliminate using traditional detection and removal tools, said Gary Golumb, principal security researcher at Netwitness, a Herndon, Va.-based vendor of security systems.

"The industry is, for many reasons, only now beginning to see the warning signs that we are not as effective as we thought we were" in dealing with security threats, Golumb said. In many cases, industry assumptions about the effectiveness of attack mitigation technologies and approaches have been "horribly off base," he added.

Security companies like RSA, Netwitness and several others argue that rather than looking to block specific threats, the better approach is to look for the telltale signs of malicious activity that malware is designed to launch. Almost all malware tools cause subtle changes in network traffic and behavior that are fairly easy to distinguish from the regular "good" traffic on a network.

The trick is to be able to effectively determine a baseline of good behavior in a way that makes it possible to filter out suspicious or malicious behavior. So, instead of looking for a Stuxnet or a Zeus or some other specific malware program, the focus should be on understanding what normal behavior is, in order to identify the abnormal or potentially malicious behavior generated by such malware.

Security incident and event management tools and network anomaly detection tools have delivered bits and pieces of this sort of capability for some time. Going forward, the goal is to integrate even more log data and other security event information from multiple sources and to correlate it using risk-based scoring methods, said Jerry Skurla, vice president of marketing at NitroSecurity, a Portsmouth, N.H.-based vendor of security incident and event management tools. "What people underestimated is the amount of data that needs to be looked at," in order to detect and effectively deal with security threats, he said.

Jaikumar Vijayan covers data security and privacy issues, financial services security and e-voting for Computerworld. Follow Jaikumar on Twitter at @jaivijayan, or subscribe to Jaikumar's RSS feed . His e-mail address is jvijayan@computerworld.com.

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