The disclosure is related to selecting and allocating one of resources in a cloud computing system to create a virtual machine. A control server may determine a resource selection reference by selecting one of fault-tolerance and network efficiency upon receipt of a request message for creating a target virtual machine. The control server may calculate at least one of a fault-tolerance point and a network efficiency point of each candidate resource according to the selected resource selection reference. The control server may select one of candidate resources to create a requested virtual machine based on at least one of the calculated fault-tolerance point and the calculated network efficiency point of the candidate resources.
