Basip a Virtual Machine Placement Technique to Reduce Energy Consumption in Cloud Data Centre

Infrastructure as a Service (IaaS) in cloud computing provides Infrastructure as a service for the demand of user from small instance to large instance in pay per use model. The services include like computer resource, networking and data storage. An API (Application Programming Interface) is used to access the infrastructure and a dashboard to control the server and to create and manage different Virtual Machines in the cloud data centre. Multiple cloud users access the service simultaneously. Due to continuous access of the services, a deadlock can happen, and it may lead to a system crash, although cloud computing is designed to overcome such problem, a proper Virtual Machine Placement Technique is needed for implementation to rule out such incidences. Deadlock can occur in cloud computing as the system is inherited from distributed computing and virtualization. A Virtual Machine Placement Technique known as BASIP is proposed to overcome the issue of deadlock by using a banker algorithm with Stochastic Integer Programming. Further, the proposed algorithm is being simulated with hundreds of servers and thousands of virtual machines. The proposed algorithm was simulated with different overload detection and VM selection algorithm. The BASIP algorithm is experimented with 800 servers with 1024 Virtual Machines. From the experimental results, BASIP algorithm reduces energy significantly. Keyword: Virtual Machine, Cloud Computing, Live Migration, Bankers and Stochastic Integer Programming, Deadlock Avoidance, Resource Allocation, Energy, Cloud Data Centre, Minimum Utilization Rank, Polynomial Regression, BASIP