Difference between revisions of "SAMM - Threat Assessment - 1"
|Line 1:||Line 1:|
<div style="float:left; width:65%;">
<div style="float:left; width:65%;">
Revision as of 19:55, 4 May 2009
| For the latest project news and information,|
join the mailing list and visit the OpenSAMM website.
Threat Assessment - 1
|Objective: Identify and understand high-level threats to the organization and individual projects|
- High-level understanding of factors that may lead to negative outcomes
- Increased awareness of threats amongst project teams
- Inventory of threats for your organization
- >50% of project stakeholders briefed on the threat models of relevant projects within past 12 months
- >75% of project stakeholders briefed on attacker profiles for relevant architectures
- Buildout and maintenance of project artifacts for threat models
- Business Owners (1 day/yr)
- Developers (1 day/yr)
- Architects (1 day/yr)
- Security Auditors (2 day/yr)
- Managers (1 day/yr)
- Strategy & Metrics - 1
- Security Requirements - 2
A. Build and maintain application-specific threat models
Based purely on the business purpose of each software project and the business risk profile (if available) identify likely worst-case scenarios for the software under development in each project team. This can be conducted using simple attack trees or through a more formal threat modeling process such as Microsoft’s STRIDE, Trike, etc.
To build attack trees, identify each worst-case scenario in one sentence and label these as the high-level goals of an attacker. From each attacker goal identified, identify preconditions that must hold in order for each goal to be realized. This information should be captured in branches underneath each goal where each branch is either a logical AND or a logical OR of the statements contained underneath. An AND branch indicates that each directly attached child nodes must be true in order to realize the parent node. An OR branch indicates that any one of the directly attached child nodes must be true in order to achieve the parent node.
Regardless of the threat modeling approach, review each current and historic functional requirement to augment the attack tree to indicate security failures relevant to each. Brainstorm by iteratively dissecting each failure scenario into all the possible ways in which an attacker might be able to reach one of the goals. After initial creation, the threat model for an application should be updated when significant changes to the software are made. This assessment should be conducted with senior developers and architects as well as one or more security auditors.
B. Develop attacker profile from software architecture
Initially, conduct an assessment to identify all likely threats to the organization based on software projects. For this assessment, consider threats to be limited to agents of malicious intent and omit other risks such as known vulnerabilities, potential weaknesses, etc.
Begin by generally considering external agents and their corresponding motivations for attack. To this list, add internal roles that could cause damage and their motivations for insider attack. Based on the architecture of the software project(s) under consideration, it can be more efficient to conduct this analysis once per architecture type instead of for each project individually since applications of architecture and business purpose will generally be susceptible to similar threats.
This assessment should be conducted with business owners and other stakeholders but also include one or more security auditors for additional perspective on threats. In the end, the goal is to have a concise list of threat agents and their corresponding motivations for attack.