Trustworthy detection and arbitration of SLA violations in the cloud

Abstract

In cloud computing, detecting violations of Service Level Agreements (SLAs) is possible by measuring certain metrics, which can be done by both the provider and the consumer of a service. However, both parties have contradicting interests with regards to these measurements, which makes it difficult to reach consensus about whether SLA violations have occurred. Within this paper, we present a solution for measuring and arbitrating SLA violations in a way that can be trusted by both parties. Furthermore, we show that this solution is not intrusive to the service and does not incur a significant overhead system load, but nevertheless provides high accuracy in detecting SLA violations.