Resource freshness and replication

• US 20070168516A1
• Filed: 02/28/2006
• Published: 07/19/2007
• Est. Priority Date: 12/05/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Suriyanarayanan, Guhan, Liu, Huisheng, Bjorner, Nikolaj
• Granted Patent: US 7,788,223 B2
• Time in Patent Office: Days
• US Class Current: 709/226
• CPC Class Codes: G06F 16/273 Asynchronous replication or reconciliation

Abstract

Aspects of the subject matter described herein relate to resource replication systems. In aspects, a mechanism is described that detects when local resources are stale. Local resources may be stale when the time between a last successful synchronization activity and a current time exceeds a staleness value. If the local resources are determined to be stale, the local member may refrain from certain synchronization activities until the local member is allowed to resume these certain synchronization activities. If the local resources are not stale, additional synchronization activities including synchronizing resources between the member and an upstream or downstream partner may be performed.

Claims

1. A computer-readable medium having computer-executable instructions, comprising:

detecting whether local resources are stale via a timestamp, wherein the timestamp indicates when a successful synchronization activity between a local machine hosting the local resources and an upstream partner of the first machine occurred, wherein the local machine and the upstream partner participate in a replica group;

if the local resources are stale, refraining from certain synchronization activities until the local machine is allowed to resume synchronization activities; and

if the local resources are not stale, performing the certain synchronization activities.

2. The computer-readable medium of claim 1, wherein detecting whether local resources are stale via a timestamp comprises determining a difference between the timestamp and a time of local system clock of the local machine.

3. The computer-readable medium of claim 1, wherein detecting that the local resources are stale comprises determining that the difference is greater than a staleness value.

4. The computer-readable medium of claim 3, wherein the staleness value is selectable by a system administrator.

5. The computer-readable medium of claim 3, wherein the staleness value is less than or equal to a tombstone period that indicates how long a resource is marked as deleted before resource metadata associated with the resource is deleted.

6. The computer-readable medium of claim 1, wherein the certain synchronization activities comprise synchronizing the local resources with resources on the upstream machine.

7. The computer-readable medium of claim 1, further comprising:

obtaining any resource on the upstream machine that is not identical to a corresponding resource on the local machine;

adding any resource that is contained on the upstream machine that is not contained on the local machine; and

marking the local resources as fresh by updating the timestamp.

8. The computer-readable medium of claim 7, further comprising deleting any resource on the local machine that is not on the upstream partner.

9. The computer-readable medium of claim 1, wherein a successful synchronization activity comprises synchronizing resources between the upstream partner and the local machine.

10. The computer-readable medium of claim 1, wherein a successful synchronization activity comprises maintaining a connection with the upstream machine.

11. A method implemented at least in part by a computer, the method comprising:

storing a value corresponding to a time of a first synchronization activity between members that replicate resources with each other;

receiving a request to engage in a second synchronization activity, wherein the request is received at a first member of the members; and

determining whether resources of the first member are fresh enough to engage in the second synchronization activity.

12. The method of claim 11, wherein the value comprises a timestamp.

13. The method of claim 11, wherein the first synchronization activity comprises a successful synchronization of resources between the first member and a partner of the first member.

14. The method of claim 11, wherein the second synchronization activity comprises a request to synchronize resources between the first member and a downstream partner of the first member.

15. The method of claim 11, wherein determining whether resources of the first member are fresh enough to engage in the second synchronization activity comprises computing a difference between the time of the first synchronization activity and a time at which the request to engage in the second synchronization is received, and if the difference is greater than a staleness value, determining that the resources of the first member are not fresh enough to engage in the second synchronization activity.

16. The method of claim 11, further comprising refusing a synchronization session associated with the request and returning a value that indicates that the resources of the first member are not fresh enough to engage in synchronization.

17. In a computing environment, an apparatus, comprising:

a resource store arranged to store resource metadata for resources that are replicated by members arranged to participate in a replica group, wherein the resource store is also arranged to store a value corresponding to a time at which a successful synchronization activity occurred; and

an update mechanism arranged to determine whether resources of the resource store are fresh enough to engage in a second synchronization activity.

18. The apparatus of claim 17, wherein the update mechanism comprises epoch logic that indicates conditions for which resources of the resource store are fresh enough to engage in the second synchronization activity.

19. The apparatus of claim 18, wherein the resources are fresh enough if the time at which a successful synchronization activity occurred is within a staleness value of a time at which the second synchronization activity is requested.