Dynamic Population Discovery for Lateral Movement Detection (Using Machine Learning)

DEF CON 24

Presented by: Rod Soto, Joseph Zadeh
Date: Saturday August 06, 2016
Time: 15:10 - 16:00
Location: Packet Hacking Village

The focus of this presentation is to describe ways to automate the discovery of different asset classes and behavioral profiles within an enterprise network. We will describe data driven techniques to derive fingerprints for specific types of individual and subgroup behaviors. The goal of these methods is to add context to communications taking place within an enterprise as well as being able to identify when certain asset profiles change there behavioral fingerprint in such a way as to indicate compromise. The type of profiles we want to discover can be tied to human behavior (User Fingerprinting) or particular asset classes like WebServers or Databases (Hardware/Software Fingerprinting). Finally enriching these profiles with a small amount of network context lets us break down the behaviors across different parts of the network topology.

These techniques become important when we want to passively monitor for certain attacks against server hardware even without visibility into the local logs running on the server. For example we will cover the automated discovery and enrichment of DMZ assets and how we use these techniques to profile when a server has been planted with a Webshell or when an asset has been used to covertly exfil data. The methods we propose should be generic to apply to a wide variety of any kind of Layer 4/ Layer 7 traffic or just PCAP data alone.

Rod Soto

Rod Soto (Twitter: @rodsoto) has over 15 years of experience in information technology and security. Currently working as a Security Researcher at Splunk User Behavioral Analytics. He has spoken at ISSA, ISC2, OWASP, DEF CON, Hackmiami, Bsides and also been featured in Rolling Stone Magazine, Pentest Magazine, Univision and CNN. Rod Soto was the winner of the 2012 Black Hat Las Vegas CTF competition and is the founder and lead developer of the Kommand & KonTroll competitive hacking Tournament series.

Joseph Zadeh

Joseph Zadeh (Twitter: @josephzadeh) studied mathematics in college and received a BS from University California, Riverside and an MS and PhD from Purdue University. While in college, he worked in a Network Operation Center focused on security and network performance baselines and during that time he spoke at DEF CON and Torcon security conferences. Most recently he joined Caspida as a security data scientist. Previously, Joseph was part of the data science consulting team at Greenplum/Pivotal helping focused on Cyber Security analytics and also part of Kaiser Permanente's first Cyber Security R&D team.


KhanFu - Mobile schedules for INFOSEC conferences.
Mobile interface | Alternate Formats