The classification unit 21 newly sets the number of servers of the base having the largest number of servers with free resources to M** (S22), and then proceeds to the process of step S25.

In step S19 of FIG. 8B, when the placement location is not taken into consideration (No), the classification unit 21 substitutes all the arrays m* in the new array m** (S23), and newly sets the number of servers M* with free resources to M** (S24), and then proceeds to the process of step S25.

That is, the classification unit 21 classifies the physical servers filtered based on the placement locations of the respective physical servers into clusters.

Steps S25 to S33 are repetitively executed on all clusters. When there are a plurality of cluster candidates, a server group is selected as a candidate in ascending order of failure probability, e.g., in ascending order of disk write count, in ascending order of operation time, or in the ascending order of temperature of the servers.

In step S26, the destination selection unit 22 determines whether or not virtual machines or containers to be newly deployed are for initial deployment. When the virtual machines or containers to be newly deployed are for initial deployment (Yes), the destination selection unit 22 calculates a total value C of the volume of free resources of the servers in this cluster n (S29).

Then, in step S30, the destination selection unit 22 determines whether or not the number of virtual machines or containers to be newly deployed is equal to or less than the total value C of the volume of free resources of the servers.

In step S30, the destination selection unit 22 proceeds to step S38 of FIG. 8C when the number of virtual machines or containers to be newly deployed is equal to or less than the total value C of the volume of free resources of the servers (Yes). Then, in step S38, the destination selection unit 22 adopts this cluster n as a deployment destination, transmits the arrays m** and the number of elements M** to the placement control unit 23, and then ends the process of FIG. 8C.

That is, in the case of initial deployment of virtual machines or containers, the destination selection unit 22 preferentially sets a cluster having a lower failure probability as the placement destination.

In step S30, the destination selection unit 22 proceeds to step S33 when the number of virtual machines or containers to be newly deployed is more than the total value C of the volume of free resources of the servers (No).

In step S26, when the virtual machines or containers to be newly deployed are not for initial deployment (No), the destination selection unit 22 proceeds to step S27 and determines whether or not the virtual machines or containers are second ones.

In step S27, when the virtual machines or containers to be newly deployed are second ones (Yes), the destination selection unit 22 determines whether or not there are virtual machines or containers of the same type that are initially deployed in this cluster n (S28). In FIG. 8B, step S28 is described as “VMs for initial deployment present in cluster n” in an abbreviated manner. When the determination in step S28 is Yes, this case corresponds to the above-mentioned deployment of second ones for redundancy or due to scale-out.

When there is no initially deployed virtual machine or container of the same type in this cluster n, the destination selection unit 22 proceeds to step S29, and sets this cluster n as a deployment target if the volume of free resources of this cluster n is acceptable.

In step S28, when there are initially deployed virtual machines or containers of the same type in this cluster n (Yes), the destination selection unit 22 proceeds to step S33, and does not set this cluster n as a deployment target.

That is, in the case of redundancy for virtual machines or containers, the destination selection unit 22 preferentially sets, as the placement destination, a cluster on which no active virtual machine or container is installed, the cluster having a lower failure probability. Therefore, the destination selection unit 22 can reduce possibility that both the active system and the standby system stop at the same time.

When the virtual machines or containers are second ones for scale-out, the destination selection unit 22 preferentially sets, as the placement destination a cluster on which no first machine or container is installed, the cluster having a lower failure probability. Therefore, the destination selection unit 22 can reduce possibility that two virtual machines or containers of the same type stop at the same time.

In step S27, when the virtual machines or containers to be newly deployed are not second ones (No), the destination selection unit 22 counts the number Nn of virtual machines or containers of the same type present in this cluster n (S31). This case corresponds to third ones or later among the above-described ones for scale-out. When the number Nn of virtual machines or containers of the same type is not 0 (No), the destination selection unit 22 proceeds to step S33, and does not set this cluster n as the deployment target.

When the number Nn of virtual machines or containers of the same type is 0 (Yes), the destination selection unit 22 proceeds to step S29, and when this cluster n is a server group that can secure resources required for the deployment, the destination selection unit 22 sets this cluster as the deployment target.

That is, in the case of scaling virtual machines or containers, the destination selection unit 22 preferentially sets, as the placement destination, a cluster on which no virtual machine or container of the same type is installed, the cluster having a lower failure probability. Therefore, the destination selection unit 22 can reduce possibility that scaled services stop at the same time.

In step S33, the destination selection unit 22 returns to step S25 when there is any unprocessed cluster, and then proceeds to step S34 when all clusters are processed.

Steps S34 to S37 are repetitively executed on all clusters.

The destination selection unit 22 calculates the total value C of the volume of free resources of the servers in this cluster n (S35).

In step S36, the destination selection unit 22 determines whether or not the number of virtual machines or containers to be newly deployed is equal to or less than the total value C of the volume of free resources of the servers. When the number of virtual machines or containers to be newly deployed is equal to or less than the total value C of the volume of free resources of the servers (Yes), the destination selection unit 22 proceeds to step S38 and adopts this cluster n as the deployment destination, and the destination selection unit 22 transmits the arrays m** and the number of elements M** to the placement control unit 23, and then ends the process of