What is threat modeling?
Threat modeling is a procedure for optimizing application, system or business process security by identifying objectives and vulnerabilities, and then defining countermeasures to prevent or mitigate the effects of threats to the system.
Threat modeling helps to identify the security requirements of a system or process -- anything that is mission-critical, processing sensitive or made up of valuable data. It is a systematic and structured process that aims to identify potential threats and vulnerabilities to reduce the risk to IT resources. It also helps IT managers understand the impact of threats, quantify their severity and implement controls.
In terms of software security, threat modeling is the most important part of software design and development. It is impossible to build applications and systems that comply with corporate security policies and privacy and regulatory requirements without evaluating and mitigating threats.
IT-based threat modeling gained traction in the 1990s with the development of threat and attacker profiles. Microsoft introduced its STRIDE (Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service, and Elevation of Privilege) threat modeling methodology in 1999. There are now many other approaches. They all involve deconstructing the elements of an application or system to identify the assets to be protected and the possible threats to be mitigated. A threat modeling methodology is a way to break down a complex process into smaller tasks making it easy to spot weaknesses.
Why is threat modeling important?
Any application or system must be designed to withstand attacks. Yet, establishing the security requirements needed to achieve this can be complex. Attackers think and act differently from developers and users.
Threat modeling is a proactive method of uncovering threats not usually considered or found through code reviews and other types of audits. It enables a project team to determine which security controls an application needs to set effective countermeasures against potential threats, and how to resolve problems early on. This approach leads to a far more secure applications, and by prioritizing anticipated threats, resources are used effectively.
Threat models are a vital part of a functional security development process. When threat modeling is part of the DevOps process developers can build security into a project during development and maintenance phases. This prevents common omissions such as failing to validate input, weak authentication, inadequate error handling and failing to encrypt data.
How does the threat modeling process work?
There are several different threat modeling frameworks and methodologies. However, the key steps are similar in most of these processes. They include:
- Form a team. This team should include all stakeholders, including business owners, developers, network architects, security experts and C-level execs. A diverse team will generate a more holistic threat model.
- Establish the scope. Define and describe what the model covers. For example, is it focused on an application, a network or the application and the infrastructure it runs on? Create an inventory of all components and data included and map them to architecture and data flow diagrams. Each data type must be classified.
- Determine likely threats. For all components that are threat targets, determine where threats exist. This what-if exercise builds broad, technical and unexpected threat scenarios, including threat or attack trees to identify possible vulnerabilities or weaknesses that could lead to compromise or failure. Threat modeling tools can help automate and streamline this step.
- Rank each threat. Determine the level of risk each threat poses and rank them to prioritize risk mitigation A simple but effective approach is to multiply the damage potential of a threat by the likelihood of it occurring.
- Implement mitigations. Decide how to mitigate each threat or reduce the risk to an acceptable level. The choices are to avoid risk, transfer it, reduce it or accept it.
- Document results. Document all findings and actions, so future changes to the application, threat landscape and operating environment can be quickly assessed and the threat model updated.
Threat modeling best practices
There are several steps to take to ensure an effective approach to threat modeling. Some of them include:
- Start early. Threat modeling can be done at any time during a project, but earlier is better as the findings can help ensure the design is secure. It is also quicker and cheaper to add security controls early in the build process.
- Get a lot of input. Soliciting input from a variety of stakeholders helps identify the widest range of potential adversaries, motives, threats, and where the most vulnerable components reside.
- Use a variety of tools. There are many tools available, including some unusual approaches. For instance, the University of Washington's Security Cards are a brainstorming tool that helps discover less common or novel attacks and how best to respond to them.
- Understand risk tolerance. Business owners in particular must fully understand and communicate their risk-tolerance levels so the correct approaches to threat mitigation can be chosen to ensure business goals are met.
- Educate everyone. Train everyone involved in the different aspects of threat modeling, so their inputs can be maximized. As always with security, threat modeling is an iterative development
Threat modeling methodologies and frameworks
Early modeling methodologies used data flow diagrams to visualize of how data moves through an application or system. However, they were too limited for modern applications that are deployed in highly interconnected environments with multiple users and devices connecting to them.
Process flow diagrams are now commonly used. They show an application or system from the perspective of user interactions and how potential attackers may try to move through the application. This makes it easier to spot and prioritize potential threats.
Attack trees are also used to visualize attacks on a system, the tree root being the goal of an attack, with the leaves being ways an attack might achieve that goal. Attack trees can be built for individual components of an application or to evaluate a specific type of attack.
Many threat modeling methodologies and frameworks have been developed. Attack-centric ones focus on the types of possible attacks, and asset-centric ones focus on the assets that need to be protected. The most common ones use the following approaches:
- Damage, Reproducibility, Exploitability, Affected users, Discoverability (DREAD) is a quantitative risk analysis that rates, compares and prioritizes a cyberthreat's severity.
- National Institute of Standards and Technology's Guide to Data-Centric System Threat Modeling focuses on protecting particular data types within systems. It models aspects of attack and defense for selected data.
- Operationally Critical Threat, Asset, and Vulnerability Evaluation (OCTAVE) provides an asset- and risk-based strategic assessment that is customizable for specific security objectives and risk management It was developed by Carnegie Mellon University for the Department of Defense.
- Process for Attack Simulation and Threat Analysis (PASTA) is a seven-step, attack-centric process designed to correlate technical requirements with business objectives, considering business impact analysis and compliance requirements.
- STRIDE is part of the Microsoft Security Development Lifecycle. It identifies system entities, events and the boundaries and then applies a set of known threats. Using it, security teams can identify potential threats.
- Trike is an open source, risk-centric methodology that ensures each asset's assigned risk level is OK with all interested parties.
- Visual, Agile, and Simple Threat (VAST) is based on ThreatModeler, an automated threat modeling tool designed to integrate into an Agile software development environment and provide actionable outputs for developers and security teams.
Find out more about threat modeling
Threat modeling tools
Threat modeling is not straightforward. There is an endless number of possible threats. Even with a small project, it makes sense to use a threat modeling tool to reduce the time and cost involved.
Threat modeling tools reduce the complexity of the process, making it structured and repeatable. That reduces the number of resources needed to create a threat model from scratch and maintain it over time. A good threat modeling tool lets users visualize, design, plan for and predict all sorts of potential threats. The key features tools should have include:
- ease of input for both system information and security rules;
- threat intelligence feed to ensure the latest identified threats are considered;
- threat dashboard with suggested mitigation strategies;
- mitigation dashboard that integrates with an issue tracker like Jira; and
- reports for compliance and stakeholders.
Some commonly used threat modeling tools include:
IriusRisk. A diagram-centric threat modeling tool with adaptive questionnaires that guide the user through the technical architecture, planned feature and security context of the application.
Microsoft Threat Modeling Tool. This free resource is designed for users who are not security experts. It provides guidance on creating and analyzing threat models as part of the Microsoft's Security Development Lifecycle. It uses of standard notation to illustrate system components, data flows and security areas, making it easy to identify classes of threats based on the structure of the software being designed.
OWASP Threat Dragon. This open source tool runs as a web or a desktop application. It records possible threats, determines mitigation approaches and shows users threat model components and threat surfaces.
SD Elements. This Security Compass tool collects and classifies system information based on vulnerabilities providing audit-ready reports.
Threagile. This is an open source integrated development environment tool that incorporates threat modeling at the application codebase. It can run using the command-line, as a Docker container or as a REST server.
ThreatModeler. Community, cloud security and application security editions automate threat modeling. They identify, predict and define threats, including prebuilt architecture templates to speed up integration.
Whichever tool is used, the threat modeling process should be repeated whenever the application, IT infrastructure or threat environment changes. That keeps the threat model current, as new threats emerge.
Analyzing threats involves time and effort. It is not a checklist exercise, but it is better to find a vulnerability and fix it before hackers discover it, and threat modeling methodology is the best way to achieve this.
Learn incident response strategies that you can use in our "Ultimate guide to cybersecurity incident response."