Energy consumption prediction based on time-series models for CPU-intensive activities in the cloud

Due to the increasing energy consumption in cloud data centers, energy saving has become a vital objective in designing the underlying cloud infrastructures. A precise energy consumption model is the foundation of many energy-saving strategies. This paper focuses on exploring the energy consumption of virtual machines running various CPU-intensive activities in the cloud server using two types of models: traditional time-series models, such as ARMA and ES, and time-series segmentation models, such as sliding windows model and bottom-up model. We have built a cloud environment using OpenStack, and conducted extensive experiments to analyze and compare the prediction accuracy of these strategies. The results indicate that the performance of ES model is better than the ARMA model in predicting the energy consumption of known activities. When predicting the energy consumption of unknown activities, sliding windows segmentation model and bottom-up segmentation model can all have satisfactory performance but the former is slightly better than the later.

Automating performance and energy consumption analysis for cloud applications

In cloud environments, IT solutions are delivered to users via shared infrastructure, enabling cloud service providers to deploy applications as services according to user QoS (Quality of Service) requirements. One consequence of this cloud model is the huge amount of energy consumption and significant carbon footprints caused by large cloud infrastructures. A key and common objective of cloud service providers is thus to develop cloud application deployment and management solutions with minimum energy consumption while guaranteeing performance and other QoS specified in Service Level Agreements (SLAs). However, finding the best deployment configuration that maximises energy efficiency while guaranteeing system performance is an extremely challenging task, which requires the evaluation of system performance and energy consumption under various workloads and deployment configurations. In order to simplify this process we have developed Stress Cloud, an automatic performance and energy consumption analysis tool for cloud applications in real-world cloud environments. Stress Cloud supports the modelling of realistic cloud application workloads, the automatic generation of load tests, and the profiling of system performance and energy consumption. We demonstrate the utility of Stress Cloud by analysing the performance and energy consumption of a cloud application under a broad range of different deployment configurations.

An effective energy testing framework for cloud workflow activities

Cloud computing as the latest computing paradigm has shown its promising future in business workflow systems facing massive concurrent user requests and complicated computing tasks. With the fast growth of cloud data centers, energy management especially energy monitoring and saving in cloud workflow systems has been attracting increasing attention. It is obvious that the energy for running a cloud workflow instance is mainly dependent on the energy for executing its workflow activities. However, existing energy management strategies mainly monitor the virtual machines instead of the workflow activities running on them, and hence it is difficult to directly monitor and optimize the energy consumption of cloud workflows. To address such an issue, in this paper, we propose an effective energy testing framework for cloud workflow activities. This framework can help to accurately test and analyze the baseline energy of physical and virtual machines in the cloud environment, and then obtain the energy consumption data of cloud workflow activities. Based on these data, we can further produce the energy consumption model and apply energy prediction strategies. Our experiments are conducted in an OpenStack based cloud computing environment. The effectiveness of our framework has been successfully verified through a detailed case study and a set of energy modelling and prediction experiments based on representative time-series models.

Interference-aware coordinated power allocation in autonomous Wi-Fi environment

Self-managed access points (APs) with growing intelligence can optimize their own performances but pose potential negative impacts on others without energy ef ciency. In this paper, we focus on modeling the coordinated interaction among interest-independent and self-con gured APs, and conduct the power allocation case study in the autonomous Wi-Fi scenario. Speci cally, we build a `coordination Wi-Fi platform (CWP), a public platform for APs interacting with each other. OpenWrt-based APs in the physical world are mapped to virtual agents (VAs) in CWP, which communicate with each other through a standard request-reply process de ned as AP talk protocol (ATP).With ATP, an active interference measurement methodology is proposed re ecting both in-range interference and hidden terminal interference, and the Nash bargaining-based power control is further formulated for interference reductions. CWP is deployed in a real of ce environment, where coordination interactions between VAs can bring a maximum 40-Mb/s throughput improvement with the Nash bargaining-based power control in the multi-AP experiments.