Malware Analysis: Investigating the Right Security Alerts

Faced with an average of 17,000 security alerts a week, security professionals play the ultimate guessing game when choosing which alerts to investigate. They can investigate an alert that proves to be a true threat and thereby shut down an attack, or they can waste valuable time investigating a security alert that proves to be a false positive, while true positives evade preventive controls. Unfortunately, history shows that security professionals have a poor track record when it comes to guessing which alerts are worth investigating.

Tracking Security Alerts

In a previous post, I wrote about how much time and money enterprises waste hunting down false positives. I cited a study Damballa commissioned from the Ponemon Institute that found enterprises waste an average of 395 hours a week chasing false positives and that, unbelievably, only 1 in 5 malware alerts deemed reliable are investigated.

While these numbers may be a bit unsettling, they don’t come as a surprise when you consider that out of the 17,000 security alerts the average enterprise receives each week, only 3,400 are relevant. The odds are stacked against security professionals, and as the many high-profile data breaches of late have demonstrated, the risk is high. It simply doesn’t make sense to throw more bodies at the problem. Even if enterprises could afford to do so, it would be a tremendous waste of resources.

Benefits of Automation

Instead, enterprises need to address malware analysis and alert investigation in the same way other parts of the business have addressed time-consuming, manual processes: with automation. Enterprises need an intelligent decision-making system (IDMS) that investigates individual security alerts to determine whether they are legitimate. The system essentially prevalidates infections, enabling teams to focus on remediation rather than investigation. But that’s just the beginning.

An IDMS can also corroborate pieces of evidence to determine with much greater confidence whether a threat exists and the severity of the risk it poses by dynamically measuring it against all other infections by device type/class (server versus laptop versus printer), by activity (data exfiltration versus click fraud) and threat intent (nuisance versus IP theft). To be clear, I am not talking about security information and event management. I’m talking about using an artificial intelligence engine to corroborate data and make a decision.

To understand the power of this, consider, for example, the same piece of evidence hitting three different devices on a network. A human would have to investigate each security alert and device separately, effectively in silos. However, an automated IDMS could investigate them simultaneously and corroborate them against every other alert in near-real time. Alone, those alerts may look fairly mundane, but when considered together and in conjunction with additional evidence, the system can determine with a high level of accuracy the risk level posed by the threat, allowing the incident response team to focus its efforts where they are deemed most beneficial.

According to the Ponemon survey, only 41 percent of respondents have automated tools to capture intelligence and evaluate the true threat caused by malware. Those that do, however, reported that an average of 60 percent of malware containment doesn’t require any human input or intervention. As a result, the people who would normally address these threats can be assigned to more strategic or proactive security projects. Meanwhile, those still dedicated to threat response can do so more efficiently and effectively because their efforts are directed at legitimate threats and the time taken to validate them has been reduced.

Manually investigating security alerts is an uphill battle that few, if any, enterprises believe they can win. Like other enterprise business processes, security alert corroboration and at least prevalidation must be automated. An intelligent decision-making system can give security professionals the advantage they need to get out of the guessing game and get into the remediation game.

Share this Article:
Brian Foster

CTO, Damballa

Brian Foster serves has Damballa’s CTO and is responsible for the company’s product vision and delivery. He brings over 25 years of successful security industry experience to Damballa. Prior to Damballa, Brian was senior vice president of product management at McAfee, where during his tenure he directed the strategy and development of consumer, mobile and enterprise security solutions. Before McAfee, Foster spent over 20 years at Symantec in a range of roles including engineering, project management, development, and product management on lines such as Norton Utilities and Norton AntiVirus. Foster earned a Bachelor of Arts degree in Economics and has attended the Executive Program in Management at the University of California, Los Angeles.