Performance Modeling and Energy Conservation in Cloud Computing

Abstract

Cloud Computing provides on-demand network access to shared computing resources to users worldwide. Data center which supports the Cloud plays an important role in the performance of Cloud. The energy consumption of data center has been an important research topic recently because of data center's high energy consumption. Many job scheduling algorithms are proposed to improve the energy efficiency while maintaining the performance of server farm. This project selects several published job assignment policies, such as E*, MEESF (Most Energy-Efficient Server First), and MAIP (most energy-efficient available server first Accounting for Idle Power), implements them with Java simulation and visual animation, and analyzes with Python program. The simulations calculate estimated statistical results of server farm under different policies and the simulated results are analyzed with Python programs to evaluate the performance of different job scheduling policies. The simulation results and analytic outcomes are also compared to check the consistence between them. This project provides a way to evaluate job assignment policies and to analyses the performance of different policies.