Threat model

Threat modeling is a process by which potential threats, such as structural vulnerabilities or the absence of appropriate safeguards, can be identified, enumerated, and mitigations can be prioritized. The purpose of threat modeling is to provide defenders with a systematic analysis of what protections need to be in play, given the nature of the system, the probable attacker’s profile, the most likely attack vectors, and the assets most desired by an attacker. Threat modeling answers questions like “Where am I most vulnerable to attack?”, “What are the most relevant threats?”, and “What do I need to do to safeguard against these threats?”.

Conceptually, most people incorporate some form of threat modeling in their daily life and don’t even realize it. Commuters use threat modeling to consider what might go wrong during the morning drive to work and to take preemptive action to avoid possible accidents. Children engage in threat modeling when determining the best path toward an intended goal while avoiding the playground bully. In a more formal sense, threat modeling has been used to prioritize military defensive preparations since antiquity.

Evolution of IT-based threat modeling

Shortly after shared computing made its debut in the early 1960s individuals began seeking ways to exploit security vulnerabilities for personal gain. As a result, engineers and computer scientists soon began developing threat modeling concepts for information technology systems.

Early IT-based threat modeling methodologies were based on the concept of architectural patterns first presented by Christopher Alexander in 1977. In 1988 Robert Barnard developed and successfully applied the first profile for an IT-system attacker.

In 1994, Edward Amoroso put forth the concept of a “threat tree” in his book, “Fundamentals of Computer Security Technology.” The concept of a threat tree was based on decision tree diagrams. Threat trees graphically represent how a potential threat to an IT system can be exploited.

Independently, similar work was conducted by the NSA and DARPA on a structured graphical representation of how specific attacks against IT-systems could be executed. The resulting representation were called “attack trees.” In 1998 Bruce Schneier published his analysis of cyber risks utilizing attack trees in his paper entitled “Toward a Secure System Engineering Methodology.” The paper proved to be a seminal contribution in the evolution of threat modeling for IT-systems. In Schneier’s analysis, the attacker’s goal is represented as a “root node,” with the potential means of reaching the goal represented as “leaf nodes.” Utilizing the attack tree in this way allowed cybersecurity professionals to systematically consider multiple attack vectors against any defined target.

In 1999, Microsoft cybersecurity professionals Loren Kohnfelder and Praerit Garg developed a model for considering attacks relevant to the Microsoft Windows development environment. (STRIDE is an acrostic for Spoofing identity, Tampering with data, Repudiation, Information disclosure, Denial of service, Elevation of privilege) The resultant mnemonic helps security professionals systematically determine how a potential attacker could utilize any threat included in STRIDE.

In 2003, OCTAVE (Operationally Critical Threat, Asset, and Vulnerability Evaluation) method, a operations-centric threat modeling methodology, was introduced with a focus on organizational risk management.

In 2004, Frank Swiderski and Window Snyder wrote “Threat Modeling,” by Microsoft press. In it they developed the concept of using threat models to create secure applications.

In 2014 Ryan Stillions expressed the idea that cyber threats should be expressed with different semantic levels, and proposed the DML (Detection Maturity Level) model. An attack is an instantiation of a threat scenario which is caused by a specific attacker with a specific goal in mind and a strategy for reaching that goal. The goal and strategy represent the highest semantic levels of the DML model. This is followed by the TTP (Tactics, Techniques and Procedures) which represent intermediate semantic levels. The lowest semantic levels of the DML model are the tools used by the attacker, host and observed network artefacts such as packets and payloads, and finally atomic indicators such as IP addresses at the lowest semantic level. Current SIEM tools typically only provide indicators at the lowest semantic levels. There is therefore a need to develop SIEM tools that can provide threat indicators at higher semantic levels.

Threat modeling methodologies for IT purposes

Conceptually, a threat modeling practice flows from a methodology. Numerous threat modeling methodologies are available for implementation. Typically, threat modeling has been implemented using one of four approaches independently, asset-centric, attacker-centric, and software-centric. Based on volume of published online content, the four methodologies discussed below are the most well known.

STRIDE methodology

The STRIDE approach to threat modeling was introduced in 1999 at Microsoft, providing a mnemonic for developers to find 'threats to our products'. STRIDE, Patterns and Practices, and Asset/entry point were amongst the threat modeling approaches developed and published by Microsoft. References to "the" Microsoft methodology commonly mean STRIDE and Data Flow Diagrams.

P.A.S.T.A.

The Process for Attack Simulation and Threat Analysis (PASTA) is a seven-step, risk-centric methodology. It provides a seven-step process for aligning business objectives and technical requirements, taking into account compliance issues and business analysis. The intent of the method is to provide a dynamic threat identification, enumeration, and scoring process. Once the threat model is completed security subject matter experts develop a detailed analysis of the identified threats. Finally, appropriate security controls can be enumerated. This methodology is intended to provide an attacker-centric view of the application and infrastructure from which defenders can develop an asset-centric mitigation strategy.

Trike

The focus of the Trike methodology is using threat models as a risk-management tool. Within this framework, threat models are used to satisfy the security auditing process. Threat models are based on a “requirements model.” The requirements model establishes the stakeholder-defined “acceptable” level of risk assigned to each asset class. Analysis of the requirements model yields a threat model from which threats are enumerated and assigned risk values. The completed threat model is used to construct a risk model based on asset, roles, actions, and calculated risk exposure.

VAST

VAST is an acronym for Visual, Agile, and Simple Threat modeling. The underlying principle of this methodology is the necessity of scaling the threat modeling process across the infrastructure and entire SDLC, and integrating it seamlessly into an Agile software development methodology. The methodology seeks to provide actionable outputs for the unique needs of various stakeholders: application architects and developers, cybersecurity personnel, and senior executives. The methodology provides a unique application and infrastructure visualization scheme such that the creation and use of threat models do not require specific security subject matter expertise.

Generally accepted IT threat modeling processes

All IT-related threat modeling processes start with creating a visual representation of the application and / or infrastructure being analyzed. The application / infrastructure is decomposed into various elements to aid in the analysis. Once completed, the visual representation is used to identify and enumerate potential threats. Further analysis of the model regarding risks associated with identified threats, prioritization of threats, and enumeration of the appropriate mitigating controls depends on the methodological basis for the threat model process being utilized.

Visual representations based on data flow diagrams

The Microsoft methodology, PASTA, and Trike each develop a visual representation of the application-infrastructure utilizing data flow diagrams (DFD). DFDs were developed in the 1970s as tool for system engineers to communicate, on a high level, how an application caused data to flow, be stored, and manipulated by the infrastructure upon which the application runs. Traditionally, DFDs utilize only four unique symbols: data flows, data stores, processes, and interactors. In the early 2000s, an additional symbol, trust boundaries, were added to allow DFDs to be utilized for threat modeling.

Once the application-infrastructure system is decomposed into its five elements, security experts consider each identified threat entry point against all known threat categories. Once the potential threats are identified, mitigating security controls can be enumerated or additional analysis can be performed.

Visual representations based on process flow diagrams

The VAST methodology distinguished between application threat models (ATM) and operational or infrastructure threat models (OTM). ATMs are built with process flow diagrams (PFD). PFDs were developed in 2011 as a tool to allow Agile software development teams to create threat models based on the application design process. Applications are decomposed into their various features or use cases. Each feature is described in terms of the widgets or code building blocks required to build that feature. Features are then connected by communication protocols. The resulting visualization is a map of how a user moves through the various features of an application.

OTMs are built with end-to-end data flow diagrams very similar to traditional DFDs. End to end data flow diagrams decompose an IT system into its various independent, grouped, and shared components. Each component is described in terms of specific attributes. Components are then connected by communication pathways and protocols.

Once the ATM or OTM is constructed, the methodology specifies how potential threats are identified, enumerated, prioritized, and associated with their relevant risks and mitigating security controls.

Threat modeling tools

There are currently a number of software tools available to help threat modeling:

• IriusRisk offers both a community and a commercial version of the tool. This tool focuses on the creation and maintenance of a live Threat Model throughout the entire SDLC. It drives the process by using fully customizable questionnaires and Risk Pattern Libraries, with flow diagramming and integration with DevSecOps (OWASP ZAP, BDD-Security, Threadfix...) to empower automation.
• Microsoft’s free threat modeling tool – the Threat Modeling Tool (formerly SDL Threat Modeling Tool). This tool also utilizes the Microsoft threat modeling methodology, is DFD-based, and identifies threats based on the STRIDE threat classification scheme. It is intended primarily for general use.
• MyAppSecurity offers a commercially available threat modeling tool - ThreatModeler It utilizes the VAST methodology, is PFD-based, and identifies threats based on a customizable comprehensive threat library. It is intended for collaborative use across all organizational stakeholders.
• PyTM is an open-source Pythonic framework for threat modeling. It encodes threat information in python code, and processes that code into a variety of forms.
• securiCAD is a threat modeling and risk management tool by the Scandinavian company foreseeti. It is intended for company cyber security management, from CISO, to security engineer, to technician. securiCAD conducts automated attack simulations to current and future IT architectures, identifies and quantifies risks holistically including structural vulnerabilities, and provides decision support based on the findings. securiCAD is offered in both commercial and community editions.
• SD Elements by Security Compass is a software security requirements management platform that includes automated threat modeling capabilities. A set of threats is generated by completing a short questionnaire about the technical details and compliance drivers of the application. Countermeasures are included in the form of actionable tasks for developers that can be tracked and managed throughout the entire SDLC.
• Tutamantic "Automated Design Analysis" is an interesting tool which provides microservices for threat modeling. In contrast to integrated tools, users upload a Visio file, and receive a spreadsheet of threats.
• OWASP Threat Dragon Project. A free, open source, online threat modeling web application including system diagramming and a rule engine to auto-generate threats/mitigations.
• Mozilla SeaSponge. A free, open source, accessible threat modeling tool from Mozilla. (Last updated in 2015)

Further fields of application

Threat modeling is being applied not only to IT but also to other areas such as vehicle, building and home automation. In this context, threats to security and privacy like information about the inhabitant's movement profiles, working times, and health situations are modeled as well as physical or network-based attacks. The latter could make use of more and more available smart building features, i.e., sensors (e.g., to spy on the inhabitant) and actuators (e.g., to unlock doors).