抄録
Cloud service rapidly becomes popular, because its feature of offering large computation resources cheaply and quickly matches with modern needs. However there are difficulties to overcome in order that it is used more widely. One of the difficulties is the unstable performance of virtual servers which are provided by cloud service. In this paper we formulate the problem of achieving guaranteed performance as an optimization problem of "virtual server - real server allocation" under constraints caused by the computation, memory, and transmission capacities of real servers and the network used in the cloud service. By solving this constrained optimization problem, we can find the allocation which has the guaranteed performance.
