Vectra Networks announced the results of the second edition of its “Post-Intrusion Report”, a real-world study about threats that evade perimeter defenses and what attackers do once they get inside your network.

Report data was collected over six-months from 40 customer and prospect networks with more than 250,000 hosts, and is compared to results in last year’s report. The new report includes detections of all phases of a cyber attack and exposes trends in malware behavior, attacker communication techniques, internal reconnaissance, lateral movement, and data exfiltration.

According to the report, there was non-linear growth in lateral movement (580%) and reconnaissance (270%) detections that outpaced the 97% increase in overall detections compared to last year. These behaviors are significant as they show signs of targeted attacks that have penetrated the security perimeter.

While command-and-control communication showed the least amount of growth (6%), high-risk Tor and external remote access detections grew significantly. In the new report, Tor detections jumped by more than 1,000% compared to last year and accounted for 14% of all command-and-control traffic, while external remote access shot up by 183% over last year.

The report is the first to study hidden tunnels without decrypting SSL traffic by applying data science to network traffic.

A comparison of hidden tunnels in encrypted traffic vs. clear traffic shows that HTTPS is favored over HTTP for hidden tunnels, indicating an attacker’s preference for encryption to hide their communications.

The increase in lateral movement and reconnaissance detections shows that attempts at pulling off targeted attacks continue to be on the rise,” said Oliver Tavakoli, Vectra Networks CTO. “The attackers’ batting average hasn’t changed much, but more at-bats invariably has translated into more hits

Key findings of the study include:

  • Botnet monetization behavior grew linearly compared to last year’s report. Ad click-fraud was the most commonly observed botnet monetization behavior, representing 85% of all botnet detections.
  • Within the category of lateral movement detections, brute-force attacks accounted for 56%, automated replication accounted for 22% and Kerberos-based attacks accounted for 16%. Although only the third most frequent detection, Kerberos-based attacks grew non-linearly by 400% compared to last year.
  • Of internal reconnaissance detections, port scans represented 53% while darknet scans represented 47%, which is fairly consistent with behavior detected last year.
  • Lateral-movement detections, which track the internal spread of malware and authentication-based attacks such as the use of stolen passwords, led the pack with over 34% of total detections.
  • Command and control detections, which identify a wide range of malicious communication techniques, were close behind with 32% of detections.
  • Botnet monetization detections track the various ways criminals make money from ad click-fraud, spamming behavior, and distributed denial of service (DDoS) attacks. These botnet-related behaviors accounted for 18% of all detections.
  • The reconnaissance category looks for internal reconnaissance performed by an attacker already inside the network and represented 13% of detections.
  • Exfiltration detections look for the actual theft of data. The good news here is that it was by far the least common category of detection at 3%.

The data in the Post-Intrusion Report is based on metadata from Vectra customers and prospects who opted to share detection metrics from their production networks. Vectra identifies active threats by monitoring network traffic on the wire in these environments. Internal host-to-host traffic and traffic to and from the Internet are monitored to ensure visibility and context of all phases of an attack.

The latest report offers a first-hand analysis of active “in situ” network threats that bypass next-generation firewalls, intrusion prevention systems, malware sandboxes, host-based security solutions, and other enterprise defenses. The study includes data from 40 organizations in education, energy, engineering, financial services, government, healthcare, legal, media, retail, services, and technology.

The full report can be found here

Advertisements