Abstract
Efficiency (in term of time consumption) and effectiveness in resources utilization are the desired quality attributes in cloud services provision. The main purpose of which is to execute jobs optimally, i.e., with minimum average waiting, turnaround and response time by using effective scheduling technique. Replication provides improved availability and scalability; decreases bandwidth use and increases fault tolerance. To speed up access, file can be replicated so a user can access a nearby replica. This paper proposes architecture to convert Globally One Cloud to Locally Many Clouds. By combining replication and scheduling, this architecture improves efficiency and easy accessibility. In the case of failure of one sub cloud or one cloud service, clients can start using another cloud under "failover" techniques. As a result, no one cloud service will go down.