Resource Manager for QoS Aware Shared Application Clusters

Nowadays, middleware technologies are largely being adopted in order to make the development of distributed enterprise applications easier and to create a level of abstraction so as to present a unified programming model to application developers. However, it is important that the middleware remains effective for such types of applications that may exhibit strict Quality of Service (QoS) requirement in terms of resource availability, reliability, scalability, and timeliness. In this paper, the system developed has a new architecture called QASAC (QoS Aware Shared Application Cluster) with SLA to allocate the resources to all clients with minimized cost of resources with the QoS metrics such as percentile response time, service availability and server utilization. The system is also augmented with PDNS to receive an update regarding failure of resource manager and route the future client request to newly elected resource manager. The above mentioned QoS guarantees can be expressed as binding contracts in terms of so called Service Level Agreement (SLA) between the execution environment and application. This work aims at extending an execution environment to make it a QoS aware middleware. The resultant middleware has the capabilities such as clustering, load balancing in order to satisfy the above mentioned QoS guarantees.