Resourceful systems for fault tolerance, reliability, and safety
ACM Computing Surveys (CSUR)
The use of executable assertions for error detection and damage assessment
Journal of Systems and Software
A Practical Approach to Programming With Assertions
IEEE Transactions on Software Engineering
ICSE '97 Proceedings of the 19th international conference on Software engineering
Integrating obstacles in goal-driven requirements engineering
Proceedings of the 20th international conference on Software engineering
Toward a resourceful method of software fault tolerance
ACM-SE 37 Proceedings of the 37th annual Southeast regional conference (CD-ROM)
Implementing Design Diversity to Achieve Fault Tolerance
IEEE Software
The contour model of block structured processes
ACM SIGPLAN Notices
xDFT: an extensible dynamic fault tolerance model for cooperative system
APWeb'06 Proceedings of the 2006 international conference on Advanced Web and Network Technologies, and Applications
Hi-index | 0.00 |
This paper examines the feasibility of creating a "resourceful" software fault-tolerance system. Current fault-tolerant methods typically replace a faulty module with a redundant backup version, making no attempt to assess and correct errors in the original module. Error-recovery options are therefore limited by the number of backup modules. In contrast, a resourceful system dynamically generates alternative error-correction strategies. Periodically, the system determines which of its pre-defined goals has not been met, then executes different strategies until its goals are achieved. We outline a resourceful fault-tolerance system that defines recovery goals and specifies separate detection and correction procedures for each goal. When errors are detected, various sequences of correction procedures are examined to identify ones that meet the recovery goals. Implementation issues such as specifying recovery goals, creating recovery options, and reducing runtime overhead are examined. We describe a strategy to increase the efficiency of our method by planning each recovery before implementing it, eliminating strategies expected to be unsuccessful, impractical, or cyclical.