An Adaptive Service Replication Framework for Managing Different Responsiveness Levels

Service replication becomes the area of interest for many providers as it can be utilized to improve both performance and availability of their own services. With the increasing number of consumers, the providers are obligated to maintain the service responsiveness (i.e. Performance and availability) with respect to the signed Service Level Agreement (SLA) between the provider and consumers. In this paper, our previous adaptive replication framework is modified to provide different responsiveness levels to available consumers. The suggested framework is calculating a specific load threshold (i.e. The time needed to initiate the adaptive service replication process) according to the response time registered within the SLA contract through applying the linear regression model. Specifically, when the current load of the running service gets over the calculated threshold, the adaptive replication process will be initiated automatically to preserve the responsiveness level that the consumer demands. The paper basically describes two different scenarios to show how the suggested framework operates for individual consumers and organizations. Finally, the experimental results demonstrate that the proposed framework prohibits the violation of response time parameter in the SLA contract.