Method and system for resource management

• US 20060190482A1
• Filed: 02/22/2005
• Published: 08/24/2006
• Est. Priority Date: 02/22/2005
• Status: Active Grant

• Current Assignee: Zhigu Holdings Limited
• Original Assignee: Microsoft Corporation
• Inventors: Kishan, Arun, Probert, David B.
• Granted Patent: US 8,091,088 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 2209/504 Resource capping
• CPC Class Codes: G06F 9/485 Task life-cycle, e.g. stopping, restarting, resuming execution G06F9/4881 takes precedence
• CPC Class Codes: G06F 9/50 Allocation of resources, e.g. of the central processing unit [CPU]
• CPC Class Codes: G06F 9/5016 the resource being the memory
• CPC Class Codes: G06F 9/5022 Mechanisms to release resources
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

The present invention manages resources in a computing device to facilitate the allocation of resources amongst competing clients operating on the device. A hierarchy of budgets is constructed to encode restrictions on the aggregated use of a resource allocated by a resource provider to one or more clients. A resource manager validates and arbitrates requests to allocate resources to the one or more clients by resource providers in accordance with the budgets comprising the hierarchy. The resource manager notifies clients of availability and shortages of resources to promote compliance with the restrictions encoded in the budgets of the hierarchy.

Claims

1. A method for facilitating the efficient management of resources in a computer system, the method comprising:

tracking values related to the use of a resource by at least one client in a first budget, the values including a first limit representing a maximum amount of the resource that can be allocated to the client and a first commit representing an amount of the resource already allocated to the client;

receiving a request by a provider of the resource to validate a client request to allocate an amount of the resource to the at least one client; and

validating the client request if increasing the first commit by the requested amount does not cause the first commit to exceed the first limit.

2. The method of claim 1, further comprising:

receiving a confirmation from the provider of the resource that the requested amount has been allocated to the at least one client; and

increasing the first commit by the requested amount.

3. The method of claim 2, wherein the first limit is soft;

and further wherein increasing the first commit by the requested amount includes increasing a normal commit up to the first limit; and

further comprising

reporting a surplus amount in excess of the first limit to the provider of the resource as an excess commit, wherein the normal commit represents an amount allocable to the at least one client at a normal service level, and the excess commit represents an amount allocable to the at least one client at a reduced service level.

4. The method of claim 1, wherein the values further include a first reservation representing a maximum amount of the resource that can be pre-allocated to the client, the first reservation being bounded by the first limit;

receiving a request by a provider of the resource to validate a client request to pre-allocate an amount of the resource to the at least one client; and

validating the client request if increasing the first commit by the requested amount will not cause the first commit to exceed the first reservation.

5. The method of claim 4, further comprising:

receiving a confirmation from the provider of the resource that the requested amount has been pre-allocated to the at least one client; and

increasing the first commit by the requested amount.

6. The method of claim 1, further comprising:

tracking values related to the use of the same resource by at least one other client in a second budget, the values including a second limit representing a maximum amount of the resource that can be allocated to the other client and a second commit representing an amount of the resource already allocated to the other client;

hierarchically relating the first and second budgets as children of a common root budget having a value of an effective limit representing a maximum amount of the resource that can be allocated to all of the clients of the first and second budgets;

receiving a request by a provider of the resource to validate a client request to allocate an amount of the resource to the at least one other client;

aggregating the first and second commits; and

validating the client request if increasing the aggregated commit by the requested amount does not cause the aggregated commit to exceed the effective limit.

7. The method of claim 1, wherein the use of the resource is further subject to a flavor restriction, the flavor restriction based on a type of the use of the resource;

and further wherein the client request is validated if the client request does not cause the client'"'"'s use of the resource to exceed the flavor restriction for the type of use of the resource.

8. The method of claim 1, further comprising:

receiving a request by the provider of the resource to arbitrate a client request for the resource;

determining a relative priority and at least one allowed action for a target client to which the resource is currently allocated; and

arbitrating in favor of the requesting client when the resource can be reclaimed from the target client based the relative priority and the at least one allowed action.

9. The method of claim 8, wherein determining the relative priority and the at least one allowed action for the target client includes generating a query to a policy database.

10. The method of claim 8, further comprising:

notifying the target client from which the resource can be reclaimed to voluntarily cede the resource.

11. The method of claim 8, wherein the at least one allowed action is one of a passive action, a pro-active action, and an aggressive action.

12. A system for managing resources in a computing device, the system comprising:

a budget hierarchy encoding restrictions on a resource;

a process operating on behalf of a client associated with a budget in the budget hierarchy;

a resource manager validating a client request for an amount of the resource based on restrictions encoded in the budget hierarchy, the restrictions including a limit representing a maximum amount of the resource that can be allocated to the client by the resource provider; and

a resource provider allocating the requested amount of the resource to the process in response to the validated client request.

13. The system of claim 12, wherein the resource manager validates the request if allocating the requested amount of the resource to the process would not cause the client to exceed their budgeted limit for the resource, and invalidates the request if allocating the requested amount of the resource to the process would cause the client to exceed their budgeted limit for the resource.

14. The system of claim 12, wherein the restrictions further include a commitment representing an amount of the resource already allocated to the client, the commitment including a normal commitment representing the amount allocated within the limit, and an excess commitment representing the amount allocated in excess of the limit, wherein the resource manager validates the client request if the requested amount added to the normal commitment is within their budgeted limit for the resource.

15. The system of claim 14, wherein the restrictions further include a reservation representing a maximum amount of the resource that can be pre-allocated to the client by the resource provider, the reservation being constrained by the limit, and further wherein the resource manager validates the client request if pre-allocating the requested amount of the resource to the process would not cause the client to exceed their budgeted reservation for the resource, and invalidates the request if pre-allocating the requested amount of the resource to the process would cause the client to exceed their budgeted reservation for the resource.

16. The system of claim 14, wherein the restrictions in the hierarchy of budgets are aggregated up the hierarchy to provide a maximum amount of the resource that can be allocated to all clients of budgets within a subtree of the hierarchy of budgets, further subject to the maximum amount of the resource that can be allocated to all clients of all budgets within the hierarchy of budgets.

17. The system of claim 13, further comprising:

a policy module providing at least one of a relative priority and an allowed action for target clients to which the resource is currently allocated; and

wherein the resource manager arbitrates the client request by arbitrating in favor of the requesting client when the resource can be reclaimed from at least one of the target clients to which the resource is currently allocated based on the at least one of the relative priority and the allowed action.

18. The system of claim 17, wherein the allowed action is one of a passive action, a pro-active action, and an aggressive action.

19. A computer-accessible medium having instructions for managing resources in a computing device, the instructions comprising:

encoding limits on allocating resources to clients in budgets organized into a budget hierarchy, the limits representing a maximum amount of the resource that can be allocated to the clients;

obtaining an active budget associated with a client making a request for a restricted resource, the active budget at a level within the budget hierarchy;

approving the request if allocating the resource to the client would not cause the client to exceed the budgeted limits in the active budget and any budgets at a same or higher level as the level of the active budget within the budget hierarchy.

20. The computer-accessible medium of claim 19, the instructions further comprising:

identifying target clients to which the resource is currently allocated;

obtaining a relative priority and allowed action set associated with each of the identified target clients; and

arbitrating the client request by reclaiming the resource from at least one of the target clients in accordance with the associated relative priorities and allowed action sets.