Optimizing run-time SOA governance through context-driven SLAs and dynamic monitoring

Abstract

End-users increasingly demand the provisioning of secure, scalable, reliable, flexible, resilient, and cost-efficient infrastructures, platforms, and software. However, the preservation of these properties, particularly in SOA and cloud environments, is extremely affected by distributed, heterogeneous, transient, and volatile context information. We envision the implementation of governance feedback loops, an innovative approach that equips service-oriented systems with run-time governance capabilities able to control the fulfillment of service level agreements (SLA) under changing execution environments. However, the effectiveness of our approach depends on the capability of governance infrastructures to guarantee the consistency between monitoring strategies, governance objectives, and context situations. To advance our vision, this paper proposes (i) contextual RDF graphs, a machine-readable specification of monitoring requirements that enable governance feedback loops with dynamic context monitoring capabilities; and (ii) context-driven SLAs, an extension of SLAs where context requirements are explicitly mapped to service level objectives (SLO) to optimize the run-time control of contracted obligations