Generic Methods for Adaptive Management of Service Level Agreements in Cloud Computing

by Edwin Yaqub
Doctoral thesis
Date of Examination:2015-10-29
Date of issue:2015-11-27
Advisor:Prof. Dr. Ramin Yahyapour
Referee:Prof. Dr. Ramin Yahyapour
Referee:Prof. Dr. Jens Grabowski
Referee:Prof. Dr. Stephan Waack
Referee:Prof. Dr. Dieter Hogrefe
Referee:Prof. Dr. Carsten Damm
Referee:Prof. Dr. Konrad Rieck
Persistent Address: http://dx.doi.org/10.53846/goediss-5387

 

 

Files in this item

Name:thesis-ey.pdf
Size:7.45Mb
Format:PDF
Description:Thesis

The following license files are associated with this item:

Abstract

English

The adoption of cloud computing to build and deliver application services has been nothing less than phenomenal. Service oriented systems are being built using disparate sources composed of web services, replicable datastores, messaging, monitoring and analytics functions and more. Clouds augment these systems with advanced features such as high availability, customer affinity and autoscaling on a fair pay-per-use cost model. The challenge lies in using the utility paradigm of cloud beyond its current exploit. Major trends show that multi-domain synergies are creating added-value service propositions. This raises two questions on autonomic behaviors, which are specifically ad- dressed by this thesis. The first question deals with mechanism design that brings the customer and provider(s) together in the procurement process. The purpose is that considering customer requirements for quality of service and other non functional properties, service dependencies need to be efficiently resolved and legally stipulated. The second question deals with effective management of cloud infrastructures such that commitments to customers are fulfilled and the infrastructure is optimally operated in accordance with provider policies. This thesis finds motivation in Service Level Agreements (SLAs) to answer these questions. The role of SLAs is explored as instruments to build and maintain trust in an economy where services are increasingly interdependent. The thesis takes a wholesome approach and develops generic methods to automate SLA lifecycle management, by identifying and solving relevant research problems. The methods afford adaptiveness in changing business landscape and can be localized through policy based controls. A thematic vision that emerges from this work is that business models, services and the delivery technology are in- dependent concepts that can be finely knitted together by SLAs. Experimental evaluations support the message of this thesis, that exploiting SLAs as foundations for market innovation and infrastructure governance indeed holds win-win opportunities for both cloud customers and cloud providers.
Keywords: Service Level Agreement (SLA); Cloud Computing; Services; SLA Negotiation; Resource Allocation; Decision Making under Uncertainty; IaaS, PaaS, SaaS
 
Statistik