On Boosting Service-oriented Dependability and Efficient Energy Usage in Cloud

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering


PhD Thesis Defence


Title: "On Boosting Service-oriented Dependability and Efficient Energy 
Usage in Cloud"

By

Mr. Subrota Kumar MONDAL


Abstract

Cloud computing systems are inherently large scale, highly distributed, 
virtualized, and operate in automated shared environments. Thus, 
performance and dependability of cloud is affected by a large number of 
parameters. Further, cloud resources need to be agile and reconfigurable 
in order to respond quickly to ever changing application demands and 
service requirements. Significant research work has been done on designing 
cloud computing infrastructures in order to boost the performance and 
dependability of cloud computing systems.

In this thesis, we develop various techniques for service-oriented 
performance and dependability evaluation. We start with a discussion on 
the modeling of evaluation techniques and metrics in different fashions 
for one of the service-oriented environments, the telecommunication 
systems. In the analysis, we focus on highlighting the factors that affect 
the performance and dependability of service. Specifically, we perform the 
service-oriented dependability analysis using an service (un)reliability 
measure called Defects Per Million (DPM) than the traditional 
system-oriented dependability analysis which do not fully reflect the 
impact of system failure-recovery behavior in service-oriented 
environments. Thereafter, we extend the aforementioned analysis to cloud 
computing systems.

A good understanding of state-of-the-art cloud infrastructures would help 
finding all potential system vulnerabilities as well as enable the cloud 
architectures providing better cloud service. Accordingly, we present an 
analysis using a framework to illuminate areas for concern and identify 
areas for potential cloud service and security enhancements. Further, 
existing cloud computing systems consume a vast amount of energy and have 
a substantial impact on the environment, consequently we present suitable 
techniques for reducing energy consumption as well as for reducing the 
environmental impact while meeting the Service Level Agreements (SLAs). We 
show the evaluation of trade-offs amongst performance, dependability, and 
sustainability as well.


Date:			Friday, 7 August 2015

Time:			1:30pm - 3:30pm

Venue:			Room 2132C
 			Lift 19

Chairman:		Prof. Bradley Foreman (PHYS)

Committee Members:	Prof. Jogesh Muppala (Supervisor)
 			Prof. Brahim Bensaou
 			Prof. Gary Chan
 			Prof. Ajay Joneja (IELM)
 			Prof. Jiannong Cao (Comp., PolyU)


**** ALL are Welcome ****