Electrical and Computer Engineering ETDs

Publication Date

9-1-2015

Abstract

Cloud Computing is rapidly becoming a ubiquitous technology. It enables an escalation in computing capacity, storage and performance without the need to invest in new infrastructure and the maintenance expenses that follow. Security is among the major concerns of organizations that are still reluctant to adopt this technology: The cloud is dynamic, and with so many different parameters involved, it is a diffi cult task to regulate it. With an approach that blends Usage Management and Statistical Learning, this research yielded a novel approach to mitigate some of the issues arising due to questionable security, and to regulate performance (utilization of resources).This research also explored how to enforce the policies related to the resources inside a Virtual Machine(VM), apart from providing initial access control. As well, this research compared various encryption schemes and observed their behavior in the cloud. We considered various components in the cloud to deduce a multi-cost function, which in turn helps to regulate the cloud. While guaranteeing security policies in the cloud, it is essential to add security to the network because the virtual cloud and SDN tie together. Enforcing network-wide policies has always been a challenging task in the domain of communication networks. Software-defined networking (SDN) enables the use of a central controller to define policies, and to use each network switch to enforce policies. While this presents an attractive operational model, it uses a very low-level framework, and is not suitable for directly implement- ing high-level policies. Therefore, we present a new framework for defining policies and easily compiling them from a user interface directly into OpenFlow actions and usage management system processes. This demonstrated capability allows cloud administrators to enforce both network and usage polices on the cloud.

Keywords

Cloud Computing, Software-Defined Networking, Usage Management

Document Type

Dissertation

Language

English

Degree Name

Computer Engineering

Level of Degree

Doctoral

Department Name

Electrical and Computer Engineering

First Advisor

Heileman, Gregory

First Committee Member (Chair)

Shu, Wei Wennie

Second Committee Member

Ghani, Nasir

Third Committee Member

Lamb, Christopher

Share

COinS