Research on tenant placement based on business relations

Abstract

Multi-tenancy supporting becomes a basic characteristic of Software-as-a-Service (SaaS) operation. In multi-tenancy environment, tenant placement problem is critical for balance the database server workload. In this paper, we discuss how the cross-tenant data access affects the workload and how to optimize resource utilization in tenant placement process. This paper comes up with a SaaS tenant data lifecycle framework to capture the placement determinant factors. According to the framework, a formal placement algorithm is presented based on a formal model of tenant data and a resource estimation method. As a result of the formalization, several suggestions for data center host to minimize the workload of database servers are presented. © 2010 IEEE.