SLA-Driven Cost-Effective Monitoring Based on Criticality for Multi-Tenant Service-Based Systems

Multi-tenant service-based systems (SBSs) have been widely used in recent years. In SBSs, Web services are composed of business processes that fulfill multiple tenants' functional and multi-dimensional quality-of-service (QoS) requirements, such as response time, throughput, and budget. QoS violations frequently occur, and the rate of service level agreement (SLA) breaches increases in a distributed and volatile operating environment. Service monitoring can be used to improve the reliability and availability of the systems by detecting runtime anomalies in a timely manner. However, this process consumes monitoring resources and incurs monitoring overhead in an SBS. In fact, different services in an SBS are not equally critical, and treating them as the same in monitoring wastes resources. Thus, it is a considerable challenge to monitor the services of a multi-tenant SBS in a cost-effective manner. In this paper, we propose SLA-driven Cost-Effective Monitoring for multi-Tenant SBSs (SCEMTS) based on criticality for formulating a monitoring strategy. In an SBS, the criticality of a service is evaluated based on two dimensions: QoS and tenants (including their priorities). To specify the different QoS requirements of tenants, tenant SLAs are used to rank tenants' priorities for weighting the criticality of a service in the monitoring resource distribution. Extensive experiments show that the SCEMTS outperforms the representative approaches in satisfying tenant SLAs, thus improving monitoring cost-effectiveness and ensuring the QoS of multi-tenant SBSs.