A Framework for Dynamic Self-optimization of Power and Dependability Requirements in Green Cloud Architectures

Abstract

I report on the activities and research challenges, their rationales, and the work in progress related to the ongoing EPSRC/UoB Bridging the Gap Fellowship project on Green Cloud Architectures. The initiative is aimed at a framework for dynamic self-optimization of cloud architectures taking into account the tradeoffs involved in maintaining acceptable dependability requirements/ Quality of Service (QoS) with minimal power at runtime. I argue that linkage between dependability requirements and power should be explicit. I motivate the need for new meters for Power-per-QoS value (and sacrifices) for cloud architectures. I motivate the need for an economics-inspired approach for dynamic self-optimization of cloud architectures. I discuss the role of Data Driven Simulation Systems in implementing such framework.