Real Time Fault Tolerant Computing for Ground-based Midcourse Defense (GMD)

Providing advanced capabilities for real-time detection, correction, and compensation of software faults can increase the long-term effectiveness of the Ground-based Midcourse Defense (GMD) system. New technologies developed to meet this challenge must ideally be targeted for deployment on highly redundant and widely distributed processing architectures, and be suitable for operational deployment to both new and legacy commercial and Government systems under evolutionary and incremental insertion schedules. This effort proposes to develop technologies for providing fault-tolerant behavior in wide-area publish/subscribe data management environments such as Command and Control Battle Management and Communication (C2BMC) systems. We propose to develop algorithms, methodologies, and mechanisms that address issues of software fault tolerance within a framework that allows us to simultaneously consider and balance the requirements for attaining goals based on additional system metrics such as performance and security. The Phase I effort will investigate and analyze technology approaches for introducing reliable fault-tolerant behavior into wide-area publish/subscribe environments. These approaches will leverage inherent beneficial characteristics of publish/subscribe environments for fault tolerance, as well as specific performance-enhancing features developed by RAM Laboratories for distributed C2BMC applications through other efforts.