Multi-tenancy in cloud computing is one of the foremost approaches to share one application instance among different customers and it is generally used by Software as a service (SaaS) providers. The main objective of the proposed work is to minimize the down time of virtual machines essential for resource provisioning using cluster based secure dynamic resource sharing. The proposed secure dynamic resource sharing approach allocates the service tenants to matched Virtual Machines(VMs) and allocates the VMs into physical host machines using the elliptic curve key based firefly optimization approach. First the functional characteristics of service users are grouped into clusters using FCM (Fuzzy C_means clustering) algorithm as tenants. After clustering, the tenant users are checked for authorization with the help of elliptic curve key value. When the users in the tenants are authorized then the grouped services are scheduled dynamically using the firefly optimization algorithm. The result of the work is appraised in terms of resource utilization, execution time, speed, and speedup.