Energy Efficiency Through Architectural Tactics for Self-Adaptive Cloud Systems

[CONTEXT] Cloud computing provides many benefits, such as virtualization, high availability, low cost of maintenance, and elasticity. These characteristics provide a flexible resource management, transparent to users, giving the illusion of access to almost unlimited resources. [PROBLEM] The increase of the energy consumption of cloud environments turns light over the amount of hardware resources provisioned in virtual machines. Self-adaptive systems are designed as a way to avoid and correct, under certain constraints, the degradation of quality of service (QoS) of software during its execution, which may be due to changes in environmental conditions. Self-adaptation brings elasticity to cloud-based applications by providing various configurations, thus avoiding the degradation of their QoS. Recent works in the literature on self-adaptive cloud systems (SACS) mainly focus on the analysis of the CPU usage and resource allocation for running cloud applications, neglecting the analysis of their impact on energy consumption. [OBJECTIVE] This PhD proposes to investigate how to integrate energy efficiency and energy awareness, as QoS criteria, in a selfadaptive architecture for cloud applications. [METHODOLOGY] It implies to study what are the necessary elements of a self-adaptive architecture of cloud applications at design-time, and how these elements should be exploited at run-time to ensure energy efficiency without impacting the performance of cloud applications.