Domain-independent monitoring and visualization of SLA metrics in multi-provider environments: (Short Paper)

Abstract

Over the past decade IT services have become increasingly multi-sourced, from IT outsourcing of functions such as network management and server provisioning to the widespread use of micro-services from different vendors in Cloud applications. Managing end-to-end service quality is becoming an increasing challenge. One of the key issues lies in the exchange of quality-related data between service participants in order to analyze not only SLA compliance but also root causes of failure. Formal SLA specification and management approaches today do not facilitate the specification of metric data in conjunction with a convenient way to supply this data to business analytics tools for detailed analysis. In this industry paper, we propose an extension of the rSLA DSL defining data items to be monitored and a specification-driven generator producing relational, denormalized time series tables accordingly. The resulting time series can be ingested by commonly used Business Intelligence tools such as Watson Analytics or Tableau in a straightforward and efficient manner. Using this approach, service integrators quickly identify problem areas in federated service delivery and consequently improve operational efficiency and avoid penalties related to SLA violations.