Hackers are Using Increasingly Sophisticated Methods of Cloaking Their Behavior When Breaching Networks
Vectra Network’s recently released Post Intrusion Report provides some good news and bad news for enterprise security teams. The good news shows evidence that companies are often detecting successful intrusions before intruders manage to exfiltrate data. The bad news is that intruders are developing new and more professional ways of hiding their presence.
Vectra Networks provides intrusion detection based on analysis of both “north-south traffic to and from the Internet as well as east-west internal traffic between network hosts.” It looks for and detects anomalous traffic behavior that indicates the presence of an intruder. This report is an analysis of metadata provided by participating customers – approximately 130 companies comprising 1.3 million hosts.
The good news stems from an initial bad start: all of the organizations showed signs of an existing and active intrusion. Vectra’s strength is in detecting the different stages in an intrusion kill chain: reconnaissance, lateral movement, and finally date exfiltration. It does this by applying scientific data analysis to traffic headers and traffic behavior.
“Of the 120 participating organizations,” notes the report, “117, or 97.5%, detected at least one of these behaviors during each month of the study.” But despite the volume of active intrusions, “The good news is exfiltration behaviors are far and away the most rarely detected phase of attack in this report. The evidence means most attacks are detected and stopped before significant damage occurs.”
The bad news, however, is that cybercriminals are responding with increasingly sophisticated methods of cloaking their behavior. “They are aware that people are starting to watch the inside of the network so they’re getting much quieter about how they’re moving laterally,” Vectra’s Wade Williamson told SecurityWeek.
Williamson highlighted two particular examples. The first is that intruders are becoming less likely to attempt brute force against privileged credentials. A good example can be found in Phineas Fisher’s hack of the Hacking Team. Once inside the network he used common and tailored tools that would not show up as a traditional anomaly (but, Williamson told SecurityWeek, probably would have been detected by internal traffic analysis) to navigate the network until he ‘found’ the credentials he needed. Brute forcing would have been detected. Stealth and patience are keywords in modern intrusions.
“Once inside the network, attackers often attempt to steal passwords or Kerberos tickets from valid users. This gives attackers access to privileged resources without additional malware or exploits and avoids attracting the attention of security teams,” warns the report. Wiiliamson told SecurityWeek that “last year, brute force was the most commonly detected lateral movement, with Kerberos attacks third. This year the results have flipped.”
The second example of increasing professionalism comes in the growing use of ‘hidden tunnels’. “We look for hidden tunnels inside of HTTP and HTTPS,” Williamson said. “In our previous report this was the 7th most detected exfiltration method; now it’s jumped to third. The intruder’s strategy is to embed small amounts of data within allowed normal conversations and protocols – he will embed some communication within an optional text field of an http packet. It doesn’t break the packet, it doesn’t make the normal web connection any different – but the destination server knows where to go to pick up the data. It’s like packet-level staganography,” he added. “This is a significant finding. It was rare in our last report, but it’s starting to move up to mainstream now. And it’s really difficult to detect this with traditional perimeter security.”
