Cost Efficient Resource Scaling Algorithm for Cloud Computing Environment

Abstract

Cloud providers have plenty of resources on their resource pool to be provided for the cloud users based on their agreement of supply from the providers' side and afford capacity from the users' side. This resource should scale up and scale down based on user demand, any cloud provider should aware of resource provisioning to be clear from resource fragmentation ,resource contention ,over provisioning and under provisioning. Most well-known cloud provider use horizontal scaling scheme for scale up resource which is take 2-5 minute to configure virtual machines, needs its own load balancer and also low resource utilization . When configuration takes place, user forced to pay extra fee for that minute. Other option is vertical scaling, resources are added on running virtual machines without shutting down the virtual machines with a virtual machines configuration time of 5 seconds which are not support on this well know cloud platform. This work therefore addressed this gap and tried to implement cost efficient resource scaling algorithm in open source cloud management software OpenStack to have an advantage of better resource utilization. We use in our experiment OpenStack platform and XEN hypervisor and we try to create 3 virtual machines .To make scale the resource we are following thresh hold value and time sires method with fixing the thresh hold value to 80% .from this experiment result we get our result is prediction error is less than 2>% and minimize monthly cost by 0.2% with a case window azure cloud provider which is better than previous related works.