Security Aware Virtual Machine Placement in Cloud

This is the era of Cloud Computing. Today Cloud providers such as Amazon Web Services (AWS), Microsoft Azure and Rackspace etc. are providing reliable, scalable, and inexpensive cloud computing services. They are providing “free to join, pay only for what you use” model which is encouraging more and more companies, Big and small to buy cloud services from them. Companies need not invest in buying infrastructure anymore instead they could rent virtual machines instances on various cloud providers like AWS, Microsoft Azure. Cloud computing is becoming a utility providing on-demand computing resource leases as transparent services to users. In a cloud data center multiple requests for sets of virtual machines need to be placed onto the physical machines or hosts. However, introducing a new virtual machine to a physical machine where a malicious virtual machine already exists could lead to potential security risks for the new virtual machine. Also cloud providers rarely give the customer any say on where their virtual machine will be placed in a cloud data center. The novelty of this research is that it focuses on virtual machine placement strategy where virtual machines are placed on physical machines based on the physical constraints defined by the user and it defines a business model which incentivizes the cloud provider for providing additional security to the cloud user. Using this placement strategy a user will be able to tell a cloud provider that he does not want his virtual machines to share space with those of his competitor or that he wants his virtual machines to be completely isolated from any other user. The research defines various placement constraints and it provides an algorithm to place virtual machines on hosts based on those constraints. It derives various conclusions on performance of this algorithm based on experiments conducted using multiple datasets of heterogeneous virtual machine requests