Multi-criteria Rank Based Geo-distributed Load Balancing for Cloud Computing

Efficient delivery of cloud computing services is subject to the efficiency of the service broker policies and the load balancing algorithms. The service broker distributes the load coming from the user base across the data centers of the Cloud Service Providers(CSPs). Selecting the appropriate data center to process the user requests is highly crucial in order to meet the required Quality-of-Service(QoS) parameters. A novel service broker policy, Multi-criteria Rank Based geo-distributed Load Balancing(MCRBgLB) algorithm is proposed that distributes the load to the data centers based on the rank that is dynamically calculated for each data center. The performance of the proposed algorithm is compared with the baseline algorithms, Closest Data Center Algorithm (CDCA) and Optimized Response Time Algorithm (ORTA) in Cloud Analyst simulation toolkit. The proposed algorithm outperforms the baseline algorithms in terms of cost by up to 27% and processing times by up to 63% in various simulation scenarios.