Applicability of the Imprecise Computation Approach for the QoS Enhancement in Distributed RTDBMS

Large amounts of real-time data is used by geographically distributed real-time applications. Using DRTDBMS1 is more and more needed to better manage the large amount of real-time data. In order to take account of unpredictable workload, Quality of Service (QoS) based approaches are the most appropriate. In the case of distributed real-time applications, it is necessary to take into account the problems of load balancing for user transactions between different nodes. Feedback Scheduling based approaches can manage the unpredictable workload variations. Further, the imprecise computation technique has been defined as a technique to enhance the Quality of Service (QoS) in centralized RTDBMS. In this paper, we propose to apply the imprecise computation approach in a distributed feedback control scheduling architecture to improve the QoS in DRTDBMS. Then, we focus on two ways of optimization in DRTDBMS, (1) the space optimization in which we propose to apply three replication data policies, and (2) the QoS optimization in which we propose an improvement of the imprecise computation to ensure a better quality of the service.