Variability in Multi-Tenant Enterprise Software

Abstract

Enterprise software applications have changed significantly over the last decades. Increasingly, software is deployed in a central location to be accessed through the internet, instead of installing software at end-users. Having software in a central location enables multi-tenancy, where multiple customers transparently share a system’s resources. Currently, multi-tenancy is a popular way to offer functionality of a software product through the internet to numerous customers, offering many advantages to both software vendors and customers. A challenge in this domain is offering variable features to multiple customers. This dissertation provides software architecture patterns for the realization of variability in multi-tenancy settings, in the domain of online enterprise software. The results support software architects on structuring the decision making process by providing a collection of multi-tenant architecture and variability patterns, guidelines for pattern selection, and a model to setup pattern evaluation and comparison sessions. The results reported have been gathered from case studies at software companies and evaluated by experts from the software industry. With these artifacts in hand, software architects can make well-informed decisions and find appropriate patterns for their specific situation, solving the challenges involved in selecting an architecture that supports multi-tenant online enterprise software. Also, these research results contribute to academia by reporting on numerous case studies in an emerging domain and presenting a vocabulary for further and more extensive research.