Performance evaluation and scheduling algorithm of transactions for mobile and real-time Databases in distributed environments

The demand on mobile and real-time database application and service in distributed environments is becoming increasingly extensive, but it is limited because the workload cannot be precisely predicted, especially possible overload. Furthermore, low bandwidth in mobile environments will lead to aborting or restarting for transactions because of competing for the limited system resources severely. In this paper, we propose a novel performance metric to evaluate performance in mobile and real-time database systems based on Quality of Service and the relationship of data items in the Directed Acyclic Graph. Imprecise computation technologies will be adapted to control system workload, and transactions will be divided into. mandatory subtransactions and optional subtransactions. Feedback loop control and Modifying On-Demand Breath-First Traversal algorithm are presented in the system to decrease miss ratio of transactions and improve data freshness. Our approach and algorithms give a robust and controlled behavior of mobile and real-time databases, even for transient overloads and with inaccurate runtime estimates of transactions. Further, simulation experiments show that the proposed algorithms can guarantee transient and steady performance not to be beyond the reference by database administrator.