Value- and debt-aware selection and composition in cloud-based service-oriented architectures using real options

H Social Sciences > HG Finance

Alzaghoul, Esra Fawaz Ahmad ; Bahsoon, Rami

University:University of Birmingham

Department:School of Computer Science

关键词: H Social Sciences;    HG Finance;   

Others  :  http://etheses.bham.ac.uk//id/eprint/6350/1/Alzaghoul15PhD.pdf

来源: University of Birmingham eTheses Repository

【 摘 要 】

This thesis presents a novel model for service selection and composition in Cloud-based Service-Oriented Architectures (CB-SOA), which is called CloudMTD, using real options, Dependency Structure Matrix (DSM) and propagation-cost metrics. CB-SOA architectures are composed of web services, which are leased or bought off the cloud marketplace. CB-SOA can improve its utility and add value to its composition by substituting its constituent services. The substitution decisions may introduce technical debt, which needs to be managed. The thesis defines the concept of technical debt for CB-SOA and reports on the available technical debt definitions and approaches in the literature. The formulation of service substitution problem and its technical debt valuation is based on options, which exploits Binomial Options Analysis. This thesis looks at different option types under uncertainty. This thesis is concerned with some scenarios that may lead to technical debt, which are related to web service selection and composition that has been driven by either a technical or a business objective. In each scenario, we are interested in three decisions (1) keep, (2) substitute or (3) abandon the current service. Each scenario takes into consideration either one or more QoS attribute dimension (e.g. Availability). We address these scenarios from an option-based perspective. Each scenario is linked to a suitable option type. A specific option type depends on the nature of the application, problem to be investigated, and the decision to be taken. In addition, we use Dependency Structure Matrix (DSM) in order to represent dependencies among web services in CB-SOA. We introduce time and complexity sensitive propagation-cost metrics to DSM to solve the problem. In addition, CloudMTD model informs the time-value of the decisions under uncertainty based on behavioral and structural aspects of CB-SOA.