We demonstrate a consensus utility accrual scheduling algorithm for distributable threads with run-time uncertainties in execution time, arrival models, and node crash failures. The DUA-CLA algorithm's message complexity (O(fn)), lower time complexity bound (O(D + fd + nk)), and failure-free execution time (O(D + nk)) are established, where D is the worst-case communication delay, d is the failure detection bound, n is the number of nodes, and f is the number of failures. The "lazy-abort" property is shown - abortion of currently-infeasible tasks is deferred until timely task completion is impossible. DUA-CLA also exhibits "schedule-safety" - threads proposed as feasible for execution by a node which fails during the decision process will not cause an otherwise-feasible thread to be excluded. These properties mark improvements over earlier strategies in common- and worst-case performance. Quantitative results obtained from our Distributed Real-Time Java implementation validate properties of the algorithm. © IFIP International Federation for Information Processing 2007.
CITATION STYLE
Anderson, J. S., Ravindran, B., & Jensen, E. D. (2007). Consensus-driven distributable thread scheduling in networked embedded systems. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4808 LNCS, pp. 247–260). Springer Verlag. https://doi.org/10.1007/978-3-540-77092-3_22
Mendeley helps you to discover research relevant for your work.