SAMM - Threat Assessment - 3

http://www.opensamm.org/downloads/BackButton.png

Results

 * Deeper consideration of full threat profile for each software project
 * Detailed mapping of assurance features to established threats against each software project
 * Artifacts to document due diligence based on business function of each software project

Add’l Success Metrics

 * >80% of project teams with updated threat models prior to every implementation cycle
 * >80% of project teams with updated inventory of third-party components prior to every release
 * >50% of all security incidents identified a priori by threat models in past 12 months

Add’l Costs

 * Project overhead from maintenance of detailed threat models and expanded attacker profiles
 * Discovery of all third-party dependencies

Add’l Personnel

 * Business Owners (1 day/yr)
 * Developers (1 day/yr)
 * Architects (1 day/yr)
 * Security Auditors (2 day/yr)
 * Managers (1 day/yr)

Related Levels

 * Security Requirements - 2 & 3

A. Explicitly evaluate risk from third-party components
Conduct an assessment of your software code-base and identify any components that are of external origin. Typically, these will include open-source projects, purchased COTS software, and online services which your software uses.

For each identified component, elaborate attacker profiles for the software project based upon potential compromise of third-party components. Based upon the newly identified attacker profiles, update software threat models to incorporate any likely risks based upon new attacker goals or capabilities.

In addition to threat scenarios, also consider ways in which vulnerabilities or design flaws in the third-party software might affect your code and design. Elaborate your threat models accordingly with the potential risks from vulnerabilities and knowledge of the updated attacker profile.

After initially conducted for a project, this must be updated and reviewed during the design phase or every development cycle. This activity should be conducted by a security auditor with relevant technical and business stakeholders.

B. Elaborate threat models with compensating controls
Conduct an assessment to formally identify factors that directly prevent preconditions for compromise represented by the threat models. These mitigating factors are the compensating controls that formally address the direct risks from software. Factors can be technical features in the software itself, but can also be process elements in the development life-cycle, infrastructure features, etc.

If using attack trees, the logical relationship represented by each branch will be either an AND or an OR. Therefore, by mitigating against just one precondition on an AND branch, the parent and all connected leaf nodes can be marked as mitigated. However, all child nodes on an OR node must be prevented before the parent can be marked as mitigated.

Regardless of threat modeling technique, identify compensating controls and annotate the threat models directly. The goal is to maximize coverage in terms of controls that mark parts of the threat model as mitigated. For any viable paths remaining, identify potential compensating controls for feedback into organizational strategy.

After initially conducted for a project, this must be updated and reviewed during the design phase or every development cycle. This activity should be conducted by a security auditor with relevant technical and business stakeholders.