Efficiency improvements for interactions of web-agents
Proceedings of the fifth international conference on Autonomous agents
Analysis and Algorithms for Restart
QEST '04 Proceedings of the The Quantitative Evaluation of Systems, First International Conference
A measurement study of the interplay between application level restart and transport protocol
ISAS'04 Proceedings of the First international conference on Service Availability
Analysis of Restart Mechanisms in Software Systems
IEEE Transactions on Software Engineering
Stochastic Models for Dependable Services
Electronic Notes in Theoretical Computer Science (ENTCS)
| Hi-index | 0.00 |
Modern complex information systems require management mechanisms that operate to a large extent independently and autonomously. One such mechanism is the restart of components or transactions in case a failure in the system occurs. In this paper we introduce a pragmatic algorithm to determine close to optimal restart times on-line. We present a method for choosing best restart times based on empirical data, if no theoretical distribution is known. The best restart time is determined based on the empirical hazard rate. We study the sample size required to come to a reasonably good estimate, the effect of the failure probability of a job and issues of parameter selection for the hazard rate estimation. The application considered in this paper is the connection setup time in HTTP GET necessary for the download of web pages.