Scheduling with Freshness and Performance Guarantees for Web Applications in the Cloud

Highly distributed data management platforms (e.g., PNUTS, Dynamo, Cassandra, and BigTable) are rapidly becoming the favorite choice for hosting modern web applications in the cloud. Among other features, these platforms rely on data partitioning, replication and relaxed consistency to achieve high levels of performance and scalability. However, these design choices often exhibit a trade-off between performance and data freshness. In this paper, in addition to performance SLAs, we also perceive an application tolerance to data staleness as another requirement determining the end-user satisfaction and our goal is to strike a fine balance between both the quality of service (QoS) and quality of data (QoD) perceived by the end-user. Towards that, we propose scheduling policies and mechanisms for efficiently allocating the recourses at each replica node so that to meet the conflicting requirements of user queries and replica updates. Our experimental results show that employing our scheduling strategies for resource allocation can provide significant improvements in the overall system utility when compared to the existing ones.