An appropriate load balancing mechanism is necessary for storage and retrieval of data using among the resources in a cloud data center to improve the efficiency. In order to balance the load, the resources need to be monitored and both the parties namely cloud customer and cloud service provider should negotiate and sign up with a Service Level Agreement (SLA) to keep smooth relationship. Hence a new Precedence Based Monitoring (PBM) algorithm is proposed in this paper for monitoring the cloud resources based on time, event and precedence of the resources.Moreover a new technique for Negotiation and SLA formation between the lead node of cloud site and consumer are proposed with intermediary agents. Both the consumer and the cloud provider are benefited by this proposed Reduced Penalty Class Algorithm (RPCA). If the user requested lease is not able to be provisioned on a particular cloud, then the intermediary agent of current cloud migrates the request of the customer to the intermediary agent of the neighboring cloud. Therefore a new mechanism for centralized load balancing among the computing nodes in a cloud site and decentralized load balancing across the cloud sites is proposed in this paper. The results obtained from the proposed system shows improvement in user satisfaction level and resource utilization of cloud provider that is the reduced due to the effective sharing of resources and efficient load balancing. © 2017. NSP Natural Sciences Publishing Cor.