Computer resource distribution method based on prediction

• US 7,500,001 B2
• Filed: 09/29/2004
• Issued: 03/03/2009
• Est. Priority Date: 11/10/2003
• Status: Expired due to Fees

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Takamoto, Yoshifumi, Tameshige, Takashi
• Primary Examiner(s): Vaughn, Jr.; William C
• Assistant Examiner(s): Cloud; Joiya M
• Application Number:
  US10/951,813
• Publication Number:
  US 20050102674A1
• Time in Patent Office: 1,616 Days
• Field of Search: 709/223, 709/224, 709/226
• US Class Current: 709/226
• CPC Class Codes: G06F 2209/5019 Workload prediction
• CPC Class Codes: G06F 9/5011 the resources being hardware resources other than CPUs, Servers and Terminals
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system

Abstract

A resource distribution method capable of lending surplus resources among a plurality of services and reducing the maintenance cost of the surplus resources is provided. Computer resources in the standby system have a dead standby state in which at least an application is not installed. A plurality of services or a plurality of users share the computer resources in the standby system. As a result, improvement of the utilization factor of idle computer resources and server integration are implemented, and the cost required to maintain the computer resources is reduced. Furthermore, load prediction is conducted as regards individual services by using past operation history. Idle computer resources secured from services having surplus and maintained are thrown in according to a result of the prediction.

Claims

1. A computer resource distribution method for a computer system in which a plurality of computer resources are assigned to a plurality of services and respective services are operated, said computer resource distribution method comprising the steps of:

specifying a period over which a load of a first service included in the plurality of services increases and a magnitude of the increasing load;

predicting a future load of other services included in the plurality of services, making a decision whether there is a margin for depriving of already assigned computer resources, and thereby selecting a second service to be deprived of already assigned computer resources; and

depriving the second service of computer resources assigned to the second service based on the period during which the load of the first service increases and the magnitude of the increasing load, and additionally assigning the deprived computer resources to the first service,

wherein the step of making a decision comprises the step of;

making a decision whether it is possible to keep a predetermined service level and execute the first service after depriving the second service of computer resources assigned to the second service.

2. A computer resource distribution method according to claim 1, further comprising the step of:

urging an administrator to input a final decision whether to execute deprivation of computer resources assigned to the second service, prior to the deprivation of computer resources.

3. A computer resource distribution method according to claim 1, further comprising the step of:

shifting assignment of computer resources from the first service to the second service at an end of the period over which the load of the first service increases.

4. A computer resource distribution method according to claim 3, wherein the shift of assignment of computer resources from the first service to the second service is done gradually over a predetermined time.

5. A computer resource distribution method for a computer system in which a plurality of computer resources are connected to a single policy server or a plurality of policy servers, said computer resource distribution method comprising the steps of:

assigning computer resources to a plurality of services and activating the computer resources as active system computer resources of respective services;

bringing a first group in remaining computer resources into a first standby state in which respective applications are installed in preparation for execution of the plurality of services; and

bringing a second group in the remaining computer resources into a second standby state in which applications for execution of the services are not installed.

6. A computer resource distribution method according to claim 5, wherein when a load of computer resources assigned to a certain service is predicted to increase, computer resources in the first standby state in preparation for execution of the service are shifted to the active system, and

wherein when the load is predicted to increase in the future, computer resources in the second standby state are shifted to the first standby state in preparation for execution of the service.

7. A computer resource distribution method according to claim 5, wherein when a load of computer resources assigned to a certain service is predicted to decrease, computer resources in the active system are shifted to the first standby state in preparation for execution of the service, and

wherein when the load is predicted to decrease in the future, computer resources in the first standby state in preparation for execution of the service are shifted to the second standby state.

8. A computer resource distribution method according to claim 5, wherein when hardware information or software information judged to cause a fault in operation of computer resources assigned to a certain service is obtained, computer resources in the first standby state in preparation for execution of the service are shifted to the active system, or computer resources in the second standby state are shifted to the first standby state in preparation for execution of the service.

9. A computer resource distribution method according to claim 5, wherein when computer resources in the first standby state in preparation for execution of the service are shifted to the active system, one or more computers in the second standby state are shifted to the first standby state in preparation for execution of the service, and

wherein when computer resources in the active system assigned to a certain service are shifted to the first standby state, one or more computers in the first standby state in preparation for execution of the service are shifted to the second standby state.

10. A computer resource distribution method according to claim 5, wherein a load or a fault in the future is predicted based on past history and external input of a certain service,

a similar prediction is conducted as regards a computer that is conducting a certain service,

computer resources in the active system or the first standby state are shifted to the second standby state, and/or computer resources in the second standby state are shifted to the active system or the first standby state, based on the predictions.