Component based energy aware multi-tenant application in software as-a service

Today as software applications tend to move towards cloud computing infrastructure, the energy requirements at datacenter also tend to rise significantly. Multi-tenancy enables SaaS providers offer a single application to multiple different tenants (i.e. organizations) simultaneously. Multi-tenancy has various advantages including effective version control, cost control and much more. However, the energy consumption at datacenter is still a challenge, in modern CPU, dynamic voltage/frequency scaling is used as a standard to achieve energy efficiency in processors. However, an idle server utilizes up to 70% of its peak energy requirement to run system software and maintain essential hardware devices. In this paper, we propose an energy-aware application component migration method that computes the load of the datacenter servers by considering the number of components attached to the servers, count of tenant users accessing the application components; components strike rate and various other vital factors in choosing components to migrate. When all components of the underutilized servers are migrated the server is powered off to conserve energy. We evaluate our proposed algorithm using discrete event simulation. Our evaluation and discussion exhibits the usefulness of the proposed solution by reducing the number of server up and reducing energy consumption at datacenter.