Application Intrusion Detection and Response Planning Methodology
In my presentation at AppSec Washington DC 2010 yesterday, I described a risk-based methodology for planning the implementation of the active defensive measures described in OWASP AppSensor.
The approach is technology agnostic but certainly needs to be tailored to an organisation's own business practices, application requirements and development & acquisition processes. I described some preliminary requirements:
- Application risk classification
- Secure coding and deployment
- Application security event logging.
With these available, I described a methodology to plan the implementation of AppSensor comprised of:
- Detection point selection
- Categorisation
- Requirements
- Model development
- Optimisation
- Code location
- Attack analysis
- Response action selection
- Strategic requirements
- Thresholds
- Model tuning
at which point the plan should be ready to implement.
The document includes some new charts and tables including:
- Composite chart of detection point categorisations
- Detection point inter-relationships
- Applicability of AppSensor detection points to application risk classification
- Detection point applicability to broad request checking and specific business logic areas
- Detection point tuning analysis considerations
- Example template for detection point specification
- Example template for a schedule of response thresholds and actions
as well as a recommendation for a baseline "quick-start" implementation.
There are also two detection point cross-references with other documents:
- attacks and weaknesses in the WASC Threat Classification v2
- risks in the OWASP Top Ten 2010
The full 80-page planning workbook can be downloaded from the OWASP web site:
I am aiming to work on additional content for this document over the next few months and have also begun devising a workshop training based on the planning workbook. The course will be aimed at system owners, architects and lead developers.
Posted on: 12 November 2010 at 11:45 hrs

Comments are filtered automatically and should appear shortly after they been checked.