A virtualization server is disclosed. The disclosed virtualization server comprises a plurality of CPUs, a plurality of GPUs, a plurality of graphic shared memories, a plurality of streaming CPUs, and a plurality of network cards, and allocates a CPU, a GPU, a graphic shared memory, a streaming CPU, and a network card to each terminal, in consideration of the number of terminals connecting to the virtualization server, the properties of applications executed by the terminals, bandwidths between the virtualization server and the terminals. Since the virtualization server can allocate a resource which matches with the performance required by each terminal, the virtualization server can be flexible to respond to an increase of the number of terminals and efficiently use resources.