Optimizing the Energy Consumption of Servers and Networks in Cloud Data Centers

Data center is a cost-effective infrastructure for storing large volumes of data and hosting large- scale service applications. Cloud computing service providers are rapidly deploying data centers across the world. with huge number of servers and switches. These data centers consume sig- nificant amounts of energy, contributing to high operational costs. Thus, optimizing the energy consumption of servers and networks in data centers can reduce operational costs. In a data center, power consumption is mainly due to servers, networking devices, and cooling systems, an effective energy saving strategy is to consolidate the computation and communication into smaller number of servers and network devices and then power off as many unneeded servers and network devices as possible. In this thesis, we propose several novel methods to reduce the energy consumption of computer systems and networks in data centers, while satisfying Quality of Service (QoS) requirements specified by cloud tenants. First, we study energy efficient scheduling of periodic real-time tasks on multi-core proces- sors with voltage islands, in which cores are partitioned into multiple blocks (termed voltage is- lands). We propose a Voltage Island Largest Capacity First (VILCF) algorithm for energy efficient scheduling of periodic real-time tasks on multi-core processors. It achieves better energy efficiency by fully utilizing the remaining capacity of an island before turning on more islands or increas- ing the voltage level of the current active islands. We provide detailed theoretical analysis of the approximation ratio of the proposed VILCF algorithm in terms of energy efficiency. Second, we study the resource allocation problem for virtual networks in data centers. A cloud tenant expresses computation requirement for each virtual machine (VM) and bandwidth