Diamond Model of Intrusion Analysis


The Diamond Model of Intrusion Analysis is a framework to analyze an intrusion event (E) by exploring the relationships among four core features: adversary, capability, infrastructure, and victim.

Diamond Model of Intrusion Analysis

Four Features

  • Adversary
    • represents the attacker
  • Infrastructure
    • refers to the tools and resources used by the adversary to carry out the intrusion
    • e.g., malware, exploit kits, C&C servers, etc.
  • Capability
    • describes the technical skills and aptitude of the adversary
      • e.g., ability to craft advanced techniques to
        • evade detection
        • exploit vulnerabilities
        • persist on target systems
  • Victim
    • represents the organization or individual targeted by adversary

How it Works

  • the four features are represented by the four vertices of a diamond shape
  • each event may also be described by meta-features
    • e.g., date/time, kill chain phase, result, etc.
  • each feature is also assigned a confidence level (C)
    • indicates data accuracy or the reliability of a conclusion or assumption assigned to the value by analysis
  • each even is defined by tuples
    • additional information about each feature is nested:
E = { {Adversary,C(adversary)},

     {Capability,C(capability)},
     
     {Infrastructure,C(infrastructure)},
     
     {Victim,C(victim)} = {
     
                       {IP,C(ip)},
                       
                       {Port,C(port)},
                       
                       {Process,C(process)}
                       
                      },
                      
     {Timestamp,C(timestamp)},
     
     { ... }
     
   }
  • power of the model lies in the ability to pivot along the vertices of the diamond to produce a complete analysis and correlation of the IoCs that represent the event
  • analyzing the four features can help to better understand the TTPs used by the adversary
    • help to develop effective strategies to defend against future intrusions
  • can identify patterns and trends to use in improving defense
  • events can be linked into attack graphs and activity threads, graphed along each vertex, representing the paths an adversary could take and those that were taken
  • threads can be assigned to activity groups
    • used to represent campaigns by particular adversaries

Attack Graph