Binary Linear Programming-based Release Planning for Multi-tenant Business SaaS

In multi-tenant Software as a Service (SaaS) business software, the degree of tenants' satisfaction is a significant indicator of the success of the SaaS system. Tenants' satisfaction can be achieved by continuously fulfilling their evolving needs. Usually, SaaS providers frequently deliver new releases of the application. Each release contains new or enhanced features. However, SaaS providers have limited resources, which makes it difficult to them to incorporate all of the tenants' requests in the next release. Therefore, some requirements shall be postponed to later releases. In order to achieve the highest possible level of tenets' satisfaction, SaaS providers shall include the most common requirements in the next release, which guarantee the satisfaction of highest possible number of tenants with less effort. Additionally, tenants' priorities and preferences about the requirements must be considered. Besides maximizing tenants' satisfaction, it is crucial to meet different types of constraints such as resource, technical, and contractual constraints. This paper identifies the factors that govern the release planning process for multi-tenant business software, which are contractual constraints, commonality of requirements, tenants' preferences and decision weights, risk, technical constraints. The first two factors are suggested by this paper, while the remaining factors are inherited from the traditional release planning process. Moreover, this paper proposes a framework that deals with the uniqueness of the release planning process in multi-tenant SaaS system. In this framework, Binary Linear Programming (BLP) is employed to optimize the selection process of the requirements that will be implemented in the next release. An experiments section is provided to illustrate the degree of satisfaction that can be achieved using the proposed framework.