Systemic Assurance


Principal Investigator:  Dr. William Scherlis, Carnegie Mellon University (scherlis@cs.cmu.edu)

Timeframe:  July 2014 to July 2016

Category: Trusted Systems


Description

Objectives: All successful resilient DoD systems require the assurance of common ilities, including safety, reliability, availability, durability, survivability, maintainability, evolvability, adaptability, and sustainability and many supporting technical attributes related to quality and security. DoD systems and environments are becoming increasingly complex and dynamic.  DoD’s net-centric cyber-physical-human systems, systems of systems (SoS) and enterprises often require changes to rapidly respond to new missions and threats.  This rapid response often requires changes in functionality, which may or may not impact key system properties.  Determining the impact on these properties often requires using “heavyweight” assurance methods, with significant cost and schedule consequences.  Improved system assurance methods will improve the ability to get rapid system changes into use with better understanding of how the system will perform.

Approach: Besides security, the engineering of resilient DoD systems requires assurance of safety, reliability, availability, durability, survivability, maintainability, evolvability, adaptability and sustainability.  All of this assurance needs to be achieved for increasingly complex, dynamic, cyber-physical-human net-centric systems and systems of systems and enterprises with needs for rapid response incompatible with most heavyweight assurance methods, process, and tools. The research program includes seven research and technology subprojects to exploit the intended synergies among the various technical approaches. The seven areas include:

  1. Develop baseline and intervention models for a selection of current standards and practices (identified in collaboration with DoD stakeholders), refining technical understanding of gaps and limitations.
  2. Undertake engineering design effort focused on integrating improved capability for traceability and other features required to support explicit modeling and management of chains of evidence.
  3. Design and implement experiments to address the challenge of rapid recertification. These include capturing evidence and assurance-related reasoning (assurance cases, models, analyses, configuration management, etc.).
  4. Develop a framework for assessment of architecture-derived quality attributes, focusing on architectural modeling and the relationship of architectural and compositional models with quality outcomes.
  5. Develop requirements elicitation and management approaches that better address quality and policy objectives.
  6. Augment and collaborate with diverse existing efforts focused on technical means to address particular quality criteria.
  7. Identify and advance areas in support of increasing automation, in order to reduce workload of developers and evaluators and to advance existing workload forward in the process, with immediate rewards.

Deliverables

Publications

Publications: None to date

Research Team

Researchers:

  • Dr. William Scherlis, Carnegie Mellon University
  • Dr. Jonathan Aldrich, Carnegie Mellon University
  • Dr. Travis D. Breaux, Carnegie Mellon University
  • Dr. David Garlan, Carnegie Mellon University
  • Dr. Christian Kästner, Carnegie Mellon University
  • Dr. Claire Le Goues, Carnegie Mellon University
  • Dr. Bradley Schmerl, Carnegie Mellon University
  • Dr. Joshua Sunshine, Carnegie Mellon University

Collaborating Institutions

Project Researchers