Energy-efficient resources management for cloud-based computing environments

Alsadie, D 2018, Energy-efficient resources management for cloud-based computing environments, Doctor of Philosophy (PhD), Science, RMIT University.

Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Alsadie.pdf Thesis application/pdf 3.21MB
Title Energy-efficient resources management for cloud-based computing environments
Author(s) Alsadie, D
Year 2018
Abstract Cloud computing is a rapidly emerging platform and has become a successful computing paradigm for offering IT related services to its customers using the Internet. Services are provided based on pay-as-you-use model and do not require any large capital investments in IT infrastructure. The benefits provided by Cloud technology are substantial leading to rapid deployment of Cloud-based infrastructure. In order to meet the substantial and growing demand for cloud computing services, cloud service provider use large numbers of strategically located large data centers. Cloud service providers have established large data centers for cloud-based services throughout the globe to provide an economical, flexible and scalable solution for their customers. These data centers use a large number of computing resources including hardware, software, and the virtual machine technologies. Such technologies allow services to be configured to meet the changing requirements of the cloud customers in terms of elasticity, scalability, and balancing of the workload. The high power consumption not only increases the operating costs of the cloud data centers but also causes substantial harm to the environment due to the release of toxic fumes by the power plant. In order to tackle the problem of large consumption of power, researchers began to focus on eliminating wasteful use of electricity by judicious utilization of computing resources. This thesis investigates two aspects—namely, sizing and effective consolidation of virtual machines—to ensure effective utilization of computing resources with the objective of decreasing the power consumption while adhering to the service level agreement with customers. The primary contributions of this study, enumerated in the thesis, are summarized below:

• A novel approach for appropriate sizing of virtual machines and allocating them to support the clusters of cloud workload with the aim of minimizing the power consumption that may be there due to improper allocation of computing resources in the cloud environment. It maps groups of tasks to customized virtual machine types. This mapping is based on the task usage patterns obtained from the analysis of the historical data extracted from utilization traces. The efficient use of computing resources on the basis of their actual requirements for a group of tasks helps to save a substantial amount of power. CHAPTER 0: LIST OF TABLES

• A new approach to optimize the migration of virtual machine to the physical machine to reduce the power consumption and reducing unwanted emissions. The proposed approach uses a correlation coefficient and predicts the future requirements of computing resources to compute the value/s of the variable accurately, and has been termed LIFE - Lowest Interdependence Factor Exponent. This variable shows the level to which a VM can be associated with a target physical machine. A higher value of LIFE will correspondingly result in a more significant impact factor influencing the performance of existing VMs whenever a VM is selected for migration to a target machine. In order to minimize performance degradation, migration of a VM to a target machine will only take place if it is found to correspond with a value of LIFE that is found to be the lowest.

• An enhancement of the proposed virtual machine placement and selection approach is by considering multiple resources with minimal effect on already running virtual machines and minimizing the migration cost in terms of multiple computing resources like bandwidth, CPU, and memory. This leads to the proposed Multiple Resourcebased VM Selection (MRVMS) approach for VM selection, and the Lowest Interdependence Factor Exponent Multiple Resources Predictive (LIFE-MP) approach for VM placement, by considering the multiple computing resources that are being used simultaneously. The MRVMS approach selects a VM with high CPU requirements and optimal memory requirement to reduce the workload of overloaded PMs with minimum migration cost. The LIFE-MP approach selects a PM at which to place the migrating VM, based on the PM with the lowest correlation coefficient value among the already-running VMs and the migrating VM to reduce performance degradation because of the VM migration.

• A novel approach to evaluate the overload condition of a physical machine that initiates migration of virtual machines with the aim of minimizing the number of migrations for reducing power waste. It presented a new proposal for the dynamic adjustment of threshold values that seeks to decrease the number of migrations in varying workload environments. The proposed approach, named the dynamic threshold-based fuzzy approach (DTFA), is a fuzzy threshold-based approach used for adjusting the threshold values of PMs in a cloud environment. The proposed method allows the number of migrations caused by overloading to be reduced and SLAs to be met. The proposed approaches were benchmarked by using Google Cloud and PlanetLab workload traces in comparison to existing respective approaches in the field.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Science
Subjects Distributed and Grid Systems
Keyword(s) Cloud computing
Power Consumption
VM Placement
Virtual machine (VM) migration
Version Filter Type
Access Statistics: 61 Abstract Views, 86 File Downloads  -  Detailed Statistics
Created: Thu, 30 May 2019, 15:40:33 EST by Keely Chapman
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us