Control server, virtual server distribution method
When plural virtual servers are distributed to plural physical servers, efficient distribution is performed in terms of the processing capacity of the physical servers and their power consumption. Firstly a second load of each virtual server in future is predicted based on a first load in a prescribed time period up to the present of each of the plural virtual servers. Next, the schedule is determined to distribute the plural virtual servers to the plural physical servers based on the second load of each virtual server so that a total of the second loads of one or a plurality of the virtual servers distributed to a physical server is within a prescribed range of proportion with respect to processing capacity of the physical server. Furthermore, the distribution is instructed (execution of redistribution) in accordance with the schedule.
1. A non-transitory recording medium storing a server control program for distributing a plurality of virtual servers to a plurality of physical servers, the server control program causing a computer to execute:
predicting a second load of each of the virtual servers in future based on a first load of each of the plurality of virtual servers in a prescribed time period up to the present;
determining a schedule to distribute the plurality of virtual servers to the plurality of physical servers based on the second load of each of the virtual servers and an initial value of a number of physical servers where the initial value is determined by a total of the second loads of the plurality of virtual servers and a product of a prescribed proportion and a total of processing capacities of the plurality of physical servers, the determining being performed so that a total of the second loads of one or more of the plurality of virtual servers distributed to an individual physical server is within an acceptable range of a proportion with respect to processing capacity of the individual physical server where the acceptable range is set by an upper limit greater than the prescribed proportion and a lower limit less than the prescribed proportion; and
issuing an instruction to distribute a portion or all of the plurality of virtual servers to a portion or all of the plurality of physical servers according to the schedule.
2. The non-transitory recording medium of claim 1 , wherein in the predicting the second load, the second load of each of the virtual servers is calculated by performing statistically processing the first load of each of the virtual servers in an operating time period of the virtual servers, from among the prescribed time period.
3. The non-transitory recording medium of claim 1 , wherein in the determining the schedule, a failed physical server of the plurality of physical servers is excluded from a target to which the plurality of virtual servers are distributed.
4. The non-transitory recording medium of claim 1 , wherein the server control program further causes the computer to issue an instruction to power off a physical server to which no virtual server is distributed as a result of distribution according to the schedule.
5. A control server to distribute a plurality of virtual servers to a plurality of physical servers, the control server comprising:
a processor configured
to predict a second load of each of the virtual servers in future based on a first load of each of the plurality of virtual servers in a prescribed time period up to the present,
to determine a schedule to distribute the plurality of virtual servers to the plurality of physical servers based on the second load of each of the virtual servers and an initial value of a number of physical servers so that a total of the second loads of one or more of the plurality of the virtual servers distributed to an individual physical server is within an acceptable range of a proportion with respect to processing capacity of the individual physical server where the initial value is determined by a total of the second loads of the plurality of virtual servers and a product of a prescribed proportion and a total of processing capacities of the plurality of physical servers, and the acceptable range is set by an upper limit greater than the prescribed proportion and a lower limit less than the prescribed proportion, and
to issue an instruction to distribute a portion or all of the plurality of virtual servers to a portion or all of the plurality of physical servers according to the schedule.
6. The control server of claim 5 , wherein the processor calculates the second load of each of the virtual servers by performing statistically processing the first load of each of the virtual servers in an operating time period of the virtual servers from among the prescribed time period.
7. The control server of claim 5 , wherein the processor excludes a failed physical server of the plurality of physical servers from a target to which the plurality of virtual servers are distributed.
8. The control server of claim 5 , wherein the processor further issues an instruction to power off a physical server to which no virtual server is distributed as a result of distribution according to the schedule.
9. A virtual server distribution method executed by a control server to distribute a plurality of virtual servers to a plurality of physical servers, the virtual server distribution method comprising:
predicting a second load of each of the virtual servers in future based on a first load of each of the plurality of virtual servers in a prescribed time period up to the present;
determining a schedule to distribute the plurality of virtual servers to the plurality of physical servers based on the second load of each of the virtual servers and an initial value of a number of physical servers where the initial value is determined by a total of the second loads of the plurality of virtual servers and a product of a prescribed proportion and a total of processing capacities of the plurality of physical servers, the determining being performed so that a total of the second loads of one or more of the plurality of virtual servers distributed to an individual physical server is within an acceptable range of a proportion with respect to processing capacity of the individual physical server where the acceptable range is set by an upper limit greater than the prescribed proportion and a lower limit less than the prescribed proportion; and
issuing an instruction to distribute a portion or all of the plurality of virtual servers to a portion or all of the plurality of physical servers according to the schedule.
10. The virtual server distribution method of claim 9 , wherein in the predicting the second load, the second load of each of the virtual servers is calculated by performing statistically processing the first load of each of the virtual servers in an operating time period of the virtual servers from among the prescribed time period.
11. The virtual server distribution method of claim 9 , wherein in the determining the schedule, a failed physical server of the plurality of physical servers is excluded from a target to which the plurality of virtual servers are distributed.
12. The virtual server distribution method of claim 9 , further comprising issuing an instruction to power off a physical server to which no virtual server is distributed as a result of distribution according to the schedule.