Scheduling Using Dynamic Priority in Real-Time Database Systems

Abstract

Real-time database systems have timing constraints associated with transactions and the database. To ensure that such a system completes as many transactions as possible without violating their timing constraints, its scheduling strategy should use information about the timing constraints associated with transactions and the database. Ideally, to enhance the predictability of the system, such a scheduling strategy should be used in all situations where there is resource contention. In this paper we describe an intelligent scheduling strategy for scheduling transactions in real-time database systems. The scheduling strategy uses additional timing information about transactions and the database to enhance the system's ability to meet transaction deadlines. In real - time database systems, if a data object is not updated for a long time, transactions that access it may consider it to be out - of - date. We introduce the concept of validity to quantify this notion of the age of a data object. We also show how to incorporate validity information in the scheduling strategy. * This work was supported in part by ONR under contract N 0014-88-K-0245 and by IBM FSD under University Agreement WG - 249153. Note: Abstract extracted from PDF file via OCR