Uniformly Distributed Load Balancing and Bandwidth-Delay Service Broker Policy in Cloud Computing

Abstract

The cloud services such as memory space, computing power, and infrastructure are provided by the cloud service provider (CSP) as per demand by the user. The performance of the cloud system depends on the load balancing and service broker policy. We proposed a technique for the load balancing in the cloud computing, which uses the ’Allocation Rank’ of the virtual machine to identify an appropriate virtual machine to serve the user’s request. In addition, we proposed a hybrid technique for the service brokering in cloud computing which use the quality of service parameters (QoS) such as ’Bandwidth’ and ’Delay’ observed in the datacenters to find an appropriate datacenter according to the user’s requirement. The proposed advance technique for the load balancing gave high utilization of all virtual machines compared to the existing load balancing policies, such as round robin load balancing policy (consider time quantum), equally spread current execution load balancing policy (consider the load status of the virtual machine), throttled load balancing policy (consider the availability status of the virtual machine), and weighted round-robin load balancing policy (based on the capacity of the virtual machine) etc. The other proposed hybrid policy for service brokering to identify an appropriate datacenter for the user requests is more efficient as compared to the existing service broker policy such as service proximity based routing policy (based on the proximity of datacenter), performance optimized routing policy (based on response time or bandwidth of datacenter), dynamically reconfigure routing policy, and enhanced proximity based routing policy (based on the proximity of the datacenter and the cost of the virtual machine).