Experts dispute sandboxing would have stopped Flame
Bigger lesson learned from Flame: Multiple layers of security needed, so when one technology fails, a second or third may succeed
By Antone Gonsalves
June 10, 2012 — CSO — At least one vendor is making the argument that sandboxing technology would have protected computer systems against Flame, but some experts are not convinced it would have caught the highly sophisticated malware package believed to be built for cyber-espionage.
Julian Waits, vice president of the Advanced Technology Group at GFI Software, argued this week in the company's blog that sandboxing would have been the backstop for antivirus software, which was unable to detect the stealthy Flame. GFI sells sandboxing technology, so the post was self-serving. Nevertheless, Waits' arguments, which other experts dispute, are worth considering in the context of a layered approach to security.
Sandbox technology runs on a virtual machine along with operating systems and business applications, watching files for unusual activity. When a suspicious file is spotted, the technology alerts security pros while logging unusual behavior, such as application changes and unusual network traffic. It is then up IT staff to decide what to do.
While sandboxing doesn't actually quarantine the file, the technology does spot threats before they can do significant damage, Waits argues. Some sandboxing technology can also generate signatures for the malware, which can be inserted in intrusion detection systems and even some anti-virus systems to prevent future infections.
Flame had evaded detection for four years, before Microsoft discovered it. Such malware can't be discovered by AV software because no signatures exist for it. "Most perimeter-based security technology can't catch it," Waits told CSO on Friday. "Those are all based on the what we know."
Despite the logic in Waits' arguments, other experts disagreed that sandboxing would have caught Flame, which Kaspersky Lab said had a command-and-control infrastructure built by people with a "world-class understanding of how to exploit software and cryptography."
"It seems to be one of the most sophisticated Windows-affecting malware, and I'm afraid sandboxing may not be effective in containing it due to the way it infects systems," Xuxian Jiang, an assistant professor in the Computer Science Department at North Carolina State University, said in an e-mail interview.
Scott Crawford, managing research director for Enterprise Management Associates, said sandboxing in general has its limitations. If applications in the sandbox-protected virtual environment have access to outside directories, file systems or other resources, then the malware can spread without detection.
"If that leads to enabling attack capabilities, then sandboxing would not be as effective, and may, in fact, be irrelevant," Crawford said by email.
One lesson learned from Flame is the need for multiple layers of security, so when one technology fails, a second or third may succeed. "How malware is designed requires a far more comprehensive approach than expanding a signature library, and compels vendors to provide much more in the way of ongoing research and analysis," 451 Research analyst Steve Coplan said.
Flame is just the latest example of how the threat landscape has changed considerably over the last few years. Hackers have gone from distributing large numbers of malware-carrying spam to targeting specific organizations with advanced techniques meant to steal high-value information. Targets typically operate in the defense industry, financial services, manufacturing, international law and government.
Experts believe Flame was built for targeted attacks against networks in the Middle East.
Read more about malware/cybercrime in CSOonline's Malware/Cybercrime section.