Consistency domains for replication in distributed computing

• US 8,639,786 B2
• Filed: 01/07/2011
• Issued: 01/28/2014
• Est. Priority Date: 01/07/2011
• Status: Active Grant

• Current Assignee: Red Hat, Inc. (International Business Machines Corporation)
• Original Assignee: Red Hat, Inc. (International Business Machines Corporation)
• Inventors: Little, Mark Cameron
• Primary Examiner(s): Chan, Wing
• Assistant Examiner(s): WOO, ANDREW M
• Application Number:
  US12/986,845
• Publication Number:
  US 20120179791A1
• Time in Patent Office: 1,117 Days
• Field of Search: None
• US Class Current: 709/221
• CPC Class Codes: G06F 9/52 Program synchronisation; Mutual exclusion, e.g. by means of semaphores

Abstract

A computing device performs an operation by a first computing device and updates a shared system state on the first computing device based on the operation. The computing device causes the shared system state to be updated on a second computing device using a strongly consistent replication mechanism, wherein the first computing device and the second computing device are members of a first consistency domain. The computing device also causes the shared system state to be updated on a third computing device using a weakly consistent replication mechanism, wherein the third computing device is a member of a second consistency domain.

Claims

1. A method comprising:

performing a first operation by a first computing device;

updating a shared system state on the first computing device in view of performing the first operation;

causing the shared system state to be updated on a second computing device using a first strongly consistent replication, wherein the first computing device and the second computing device are members of a first consistency domain, wherein members of the first consistency domain use the first strongly consistent replication to update the shared system state at other members of the first consistency domain; and

causing the shared system state to be updated on a third computing device using a weakly consistent replication, wherein the third computing device is a member of a second consistency domain, wherein members of the second consistency domain use a second strongly consistent replication to update the shared system state at other members of the second consistency domain, wherein the members of the first consistency domain use the weakly consistent replication to update the shared system state at the members of the second consistency domain, wherein the second consistency domain is inconsistent from the first consistency domain by at most a skew value that is defined in view of a time difference, and wherein members of the second consistency domain achieve a consistent state with members of the first consistency domain within the time difference.

2. The method of claim 1, wherein:

causing the shared system state to be updated on the second computing device comprises sending a first message to the second computing device using the first strongly consistent replication, wherein the first message causes the second computing device to update the shared system state on the second computing device; and

causing the shared system state to be updated on the third computing device comprises sending a second message to the third computing device using the weakly consistent replication, wherein the second message causes the third computing device to update the shared system state on the third computing device.

3. The method of claim 2, further comprising:

after receiving a response message from the second computing device indicating that the second computing device has updated the shared system state, performing a second operation by the first computing device and updating the shared system state on the first computing device in view of performing the second operation, regardless of whether the third computing device has updated the shared system state.

4. The method of claim 1, wherein the weakly consistent replication includes a gossip communication protocol.

5. A non-transitory computer readable storage medium including instructions that, when executed by a first computing device, cause the first computing device to perform operations comprising:

performing a first operation by the first computing device;

updating a shared system state on the first computing device in view of performing the first operation;

causing the shared system state to be updated on a second computing device using a first strongly consistent replication, wherein the first computing device and the second computing device are members of a first consistency domain, wherein members of the first consistency domain use the first strongly consistent replication to update the shared system state at other members of the first consistency domain; and

causing the shared system state to be updated on a third computing device using a weakly consistent replication, wherein the third computing device is a member of a second consistency domain, wherein members of the second consistency domain use a second strongly consistent replication to update the shared system state at other members of the second consistency domain, wherein the members of the first consistency domain use the weakly consistent replication to update the shared system state at the members of the second consistency domain, wherein the second consistency domain is inconsistent from the first consistency domain by at most a skew value that is defined in view of a time difference, and wherein members of the second consistency domain achieve a consistent state with members of the first consistency domain within the time difference.

6. The non-transitory computer readable storage medium of claim 5, wherein:

causing the shared system state to be updated on the second computing device comprises sending a first message to the second computing device using the first strongly consistent replication, wherein the first message causes the second computing device to update the shared system state on the second computing device; and

causing the shared system state to be updated on the third computing device comprises sending a second message to the third computing device using the weakly consistent replication, wherein the second message causes the third computing device to update the shared system state on the third computing device.

7. The non-transitory computer readable storage medium of claim 6, wherein the operations further comprise:

after receiving a response message from the second computing device indicating that the second computing device has updated the shared system state, performing a second operation by the first computing device and updating the shared system state on the first computing device in view of performing the second operation, regardless of whether the third computing device has updated the shared system state.

8. The non-transitory computer readable storage medium of claim 5, wherein the weakly consistent replication includes a gossip communication protocol.

9. A system comprising:

a memory to store instructions; and

a first computing device to execute the instructions to;

perform a first operation;

update a shared system state on the first computing device in view of the performance of the first operation;

cause the shared system state to be updated on a second computing device using a first strongly consistent replication, wherein the first computing device and the second computing device are members of a first consistency domain, wherein members of the first consistency domain use the first strongly consistent replication to update the shared system state at other members of the first consistency domain; and

cause the shared system state to be updated on a third computing device using a weakly consistent replication, wherein the third computing device is a member of a second consistency domain, wherein members of the second consistency domain use a second strongly consistent replication to update the shared system state at other members of the second consistency domain, wherein the members of the first consistency domain use the weakly consistent replication to update the shared system state at the members of the second consistency domain, wherein the second consistency domain is inconsistent from the first consistency domain by at most a skew value that is defined in view of a time difference, and wherein members of the second consistency domain achieve a consistent state with members of the first consistency domain within the time difference.

10. The system of claim 9, wherein:

the second computing device, which is networked with the first computing device, is to receive a first message from the first computing device via the first strongly consistent replication, to process the first message to update the shared system state on the second computing device, and to send a response message to the first computing device indicating that the second computing device has updated the shared system state; and

the third computing device, which is networked with the first computing device, is to receive a second message from the first computing device via the weakly consistent replication, and to process the second message to update the shared system state on the third computing device.

11. The system of claim 10, wherein the first computing device is further to execute the instructions to:

after receiving a response message from the second computing device indicating that the second computing device has updated the shared system state, perform a second operation and update the shared system state on the first computing device in view of performing the second operation, regardless of whether the third computing device has updated the shared system state.

12. The system of claim 9, wherein the weakly consistent replication includes a gossip communication protocol.