An adaptive scheduler for distributed real-time database systems

Abstract

In a distributed real-time database system, a remote transaction usually accesses some data located at remote sites. However, traditional real-time schedulers do not consider the impact of communication delay in transferring remote data, resulting in a high MissRatio for remote transactions. This study proposes a new real-time scheduler called flexible high reward (FHR) to reduce the MissRatio and lessen the unnecessary waiting time due to communication delay for a remote transaction. This scheduler gives a remote transaction a slightly higher priority and adjusts a remote transaction’s waiting status such that the resources held by a waiting remote transaction can be released as soon as possible. Through the underlying operating system, local transactions can also be executed while a remote transaction is waiting for remote data. Simulation results demonstrate that the FHR outperforms current real-time schedulers such as earliest deadline, highest value, and highest reward and urgency.