Deadline missing predictor based on aperiodic server queue length for distributed systems

Authors:
P. D. M. Plentz;C. Montez;R. S. de Oliveira
Affiliations:
Programa de Pós-Graduação em Engenharia Elétrica, Federal University of Santa Catarina, 88040-900 Florianópolis, SC, Brazil;Programa de Pós-Graduação em Engenharia Elétrica, Federal University of Santa Catarina, 88040-900 Florianópolis, SC, Brazil;Programa de Pós-Graduação em Engenharia Elétrica, Federal University of Santa Catarina, 88040-900 Florianópolis, SC, Brazil
Venue:
Computer Communications
Year:
2008

Citing 8
Cited 0

Deadline Assignment in a Distributed Soft Real-Time System

IEEE Transactions on Parallel and Distributed Systems
Scheduling Distributable Real-Time Threads in Tempus Middleware

ICPADS '04 Proceedings of the Parallel and Distributed Systems, Tenth International Conference
Predicting application run times with historical information

Journal of Parallel and Distributed Computing
Portable and efficient distributed threads for Java

Proceedings of the 5th ACM/IFIP/USENIX international conference on Middleware
A Real-Time Performance Comparison of Distributable Threads and Event Channels

RTAS '05 Proceedings of the 11th IEEE Real Time on Embedded Technology and Applications Symposium
Application performance prediction in autonomic systems

Proceedings of the 44th annual Southeast regional conference
Soft real-time task response time prediction in dynamic embedded systems

SEUS'07 Proceedings of the 5th IFIP WG 10.2 international conference on Software technologies for embedded and ubiquitous systems
Analysis of deadline assignment methods in distributed real-time systems

Computer Communications

Quantified Score

Hi-index	0.24

Visualization

Abstract

In this paper, a deadline missing prediction mechanism based on the aperiodic server queue length (ASQ) is presented. This mechanism uses the server queue length and the past response times of a distributed thread to define the probability of it meeting its deadline. This mechanism uses linear regression and it considers the probabilistic knowledge of the conditional paths that a distributed thread can follow when executing remote calls. Historical data hosted in the source node of a distributed thread supply the information used by the prediction mechanism to define the probability of a distributed thread to meet its deadline.