Evaluating virtual machine allocations in clouds

Abstract

Cloud Services provides virtual machines (VM) to millions of users with thousands of different configurations. To do this effectively requires the use of bin packing algorithm. This project aims to analyse and optimise pre-existing bin packing algorithms and find out which algorithm would be the best for allocating virtual machines. The tested algorithm was First Fit, Next Fit, Best Fit, and Worst Fit. Code is written in Python and tested on Azure public dataset for bin packing. The data output of the written code was analysed via chart visualisation. This project aims to better understand the implementation of bin packing algorithm on virtual machine allocation.