Implementation of locust inspired scheduling algorithm with huge number of servers for energy efficiency in a cloud datacenter

In recent times, with the rapid development of cloud computing has affected to energy consumption which gives negative impact towards the environment through production of carbon dioxide. A decentralized Locust-inspired scheduling algorithm (LACE) is one way to minimize the level of energy consumption in cloud datacenters. LACE algorithm is used to schedule and optimize Virtual Machine (VMs) allocation across the servers according to behaviour obtained from locust nature. LACE migrate the VM from under loaded server to other overloaded server in order to decrease the total number of running server. The running servers can be shut down and save the energy used. In the benchmark paper, the result of implementation of LACE algorithm in 400, 600, 800 and 1000 servers were plotted at different graphs. No comparison between the results has been made into one graph. Moreover, the implementation of LACE algorithm in datacenter consisting of 400, 600, 800 and 1000 servers only were created. It does not consider the LACE algorithm implemented in huge number of server in one Cloud datacenter. So, the objective for this paper is to evaluate the results in the benchmark paper and to evaluate the implementation of LACE algorithm in a huge number of servers within one cloud datacenter. LACE algorithm is executed in 1000, 2000, 3000 and 4000 servers to see the performance in Cloud datacenter. Cloudsim is used as Discrete Event Simulation tool and Java as coding language to evaluate LACE algorithm. The performance metrics are measured is energy consumption. The result show that if the request is high, the amount of energy consumption decrease because more number of migrations occur and more running servers used can be shut down. At low request, there is no any significance effect the level of energy consumption between the distinct number of servers since less number of migration occur.

Text FSKTM 2019 35 IR.pdf Download (1MB)

Actions (login required)

View Item