Technical debt-aware and evolutionary adaptation for service composition in SaaS clouds

Kumar, Satish (2021). Technical debt-aware and evolutionary adaptation for service composition in SaaS clouds. University of Birmingham. Ph.D.

Preview

Kumar2021PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (3MB) | Preview

Abstract

The advantages of composing and delivering software applications in the Cloud-Based Software as a Service (SaaS) model are offering cost-effective solutions with minimal resource management. However, several functionally-equivalent web services with diverse Quality of Service (QoS) values have emerged in the SaaS cloud, and the tenant-specific requirements tend to lead the difficulties to select the suitable web services for composing the software application. Moreover, given the changing workload from the tenants, it is not uncommon for a service composition running in the multi-tenant SaaS cloud to encounter under-utilisation and over-utilisation on the component services that affects the service revenue and violates the service level agreement respectively. All those bring challenging decision-making tasks: (i) when to recompose the composite service? (ii) how to select new component services for the composition that maximise the service utility over time? at the same time, low operation cost of the service composition is desirable in the SaaS cloud. In this context, this thesis contributes an economic-driven service composition framework to address the above challenges. The framework takes advantage of the principal of technical debt- a well-known software engineering concept, evolutionary algorithm and time-series forecasting method to predictively handle the service provider constraints and SaaS dynamics for creating added values in the service composition. We emulate the SaaS environment setting for conducting several experiments using an e-commerce system, realistic datasets and workload trace. Further, we evaluate the framework by comparing it with other state-of-the-art approaches based on diverse quality metrics.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):