ABSTRACT
Cloud computing and big data has led to the multi-cloud architecture that enables multiple applications to be hosted and executed on various servers that are either near or remote. In order to support multiple executions, there exists a large number of servers hosted on data-centres that are interconnected by high bandwidth networks. The resources in the server are shared by a number of Virtual Machines (VMs) that facilitates the customer oriented services. The data-centres that host multiple VMs, results in high network traffic. This traffic is mainly due to the applications hosted on the VMs. Hence, it is highly required to identify the type of the application hosted on the VM and classify the servers accordingly. The server classification enables traffic to be monitored and handled in an optimized way. Thus to address this VM classification, this paper proposes an Modified Extreme Learning Machines to classify the type of VMs hosted by the servers and the jobs are scheduled using Modified Weighted Round Robin scheduling algorithm.
