Behavioral Analysis of Multitenant SaaS Applications

Abstract

Cloud Computing has received significant attention as it has changed the way of computation. Looking at the benefits of a cloud, more demand for cloud computing services and Multitenancy. Multitenancy offers various advantages including low infrastructure cost, low maintenance, and reducing the application hosting cost. This paper proposes a Multitenant Multiservice SaaS model for designing Multitenant SaaS applications. The model provides a configurable service accessed by different Tenants. The paper further evaluates the Performance implementing this model using data layer patterns as Tenant Specific and Tenant Shared and generates results using Apache JMeter. Our results show that under the Tenant Specific as a Single tenant, Response time is better than Tenant Shared for increasing load defined by the number of concurrent users. With Tenant Shared schema handling multiple tenants with a single instance, the Response time with an increasing number of tenants and concurrent users is almost stable. Since the Tenant Shared model is resource efficient as compared to Tenant Specific, its performance is affected due to shared instances instead of separate application instances for each tenant.