SLA-driven Load Scheduling in Multi-tier Cloud Computing: Financial Impact Considerations

A cloud service provider strives to effectively provide a high Quality of Service (QoS) to client jobs. Such jobs vary in computational and Service-Level-Agreement (SLA) obligations, as well as differ with respect to tolerating delays and SLA violations. The job scheduling plays a critical role in servicing cloud demands by allocating appropriate resources to execute client jobs. The response to such jobs is optimized by the cloud service provider on a multi-tier cloud computing environment. Typically, the complex and dynamic nature of multi-tier environments incurs difficulties in meeting such demands, because tiers are dependent on each others which in turn makes bottlenecks of a tier shift to escalate in subsequent tiers. However, the optimization process of existing approaches produces single-tier-driven schedules that do not employ the differential impact of SLA violations in executing client jobs. Furthermore, the impact of schedules optimized at the tier level on the performance of schedules formulated in subsequent tiers tends to be ignored, resulting in a less than optimal performance when measured at the multi-tier level. Thus, failing in committing job obligations incurs SLA penalties that often take the form of either financial compensations, or losing future interests and motivations of unsatisfied clients in the service provided. Therefore, tolerating the risk of such delays on the operational performance of a cloud service provider is vital to meet SLA expectations and mitigate their associated commercial penalties. Such situations demand the cloud service provider to employ scalable service mechanisms that efficiently manage the execution of resource loads in accordance to their financial influence on the system performance, so as to ensure system reliability and cost reduction. In this paper, a scheduling and allocation approach is proposed to formulate schedules that account for differential impacts of SLA violation penalties and, thus, produce schedules that are optimal in financial performance. A queue virtualization scheme is designed to facilitate the formulation of optimal schedules at the tier and multi-tier levels of the cloud environment. Because the scheduling problem is NPhard, a biologically inspired approach is proposed to mitigate the complexity of finding optimal schedules. The reported results in this paper demonstrate the efficacy of the proposed approach in formulating costoptimal schedules that reduce SLA penalties of jobs at various architectural granularities of the multi-tier cloud environment.