Rapid provisioning of virtual machines based on multi-dimensional user request patterns in a cloud

摘要

A method for pre-provisioning a virtual machine instance includes deriving a weight function for at least one virtual machine request based on historical data of the at least one virtual machine request to create a weighted virtual machine request for each request, capturing at least one inter-arrival pattern of requests, and applying the at least one pattern to the at least one weighted request to create a prioritized ordering of the at least one weighted request, and pre-provisioning at least one instance of a virtual machine image based on the prioritized ordering of the at least one weighted request.