Threat modeling is the practice of systematically identifying and evaluating threats to a system. Several frameworks exist to guide this process, each with different strengths. This guide compares the most widely used approaches.
Best for: Developer teams doing design-time threat modeling
STRIDE (Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service, Elevation of Privilege) is Microsoft's framework for categorizing threats. It's most useful during the design phase and maps naturally to security controls.
Best for: Risk-focused organizations that need to tie threats to business impact
Process for Attack Simulation and Threat Analysis (PASTA) is a risk-centric framework that ties technical threats to business objectives. It's more complex than STRIDE but produces more actionable risk prioritization.
Best for: Systems that handle personal data and require privacy threat modeling
LINDDUN focuses on privacy threats: Linkability, Identifiability, Non-repudiation, Detectability, Disclosure of information, Unawareness, and Non-compliance.
Best for: Operations-focused teams that need to map threats to real adversary TTPs
Using MITRE ATT&CK as a threat modeling foundation produces threat models that are directly actionable for detection engineering and red team planning.
In practice, combining frameworks produces the best results: STRIDE for design-time threat identification, PASTA for risk prioritization, and ATT&CK for operational planning.