Cross-region replication architecture

• US 10,521,449 B1
• Filed: 12/17/2014
• Issued: 12/31/2019
• Est. Priority Date: 12/17/2014
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Schwartz, Benjamin Aldouby, Shams, Khawaja Salman, Shami, Usman Ahmed, Yanacek, David Craig, Tran, Khai Quang
• Primary Examiner(s): Robinson, Greta L
• Application Number:
  US14/573,956
• Time in Patent Office: 1,840 Days
• Field of Search: 707610, 707625, 707628, 707631, 707638, 707802, 707803
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• CPC Class Codes: H04L 67/1095 Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes

Abstract

One or more computing nodes located in a first region may maintain a first collection of data items. A second set of one or more computing nodes in a second region may maintain a collection of data items that is at least a partial replica of the first collection. Notifications of changes to the first collection may be transmitted, via a broadcast communications channel, to a replication module, which may be included in a client application. The replication module may transmit requests to update the second collection of data based on changes to the first collection. Conflicts may be resolved through a last-write wins policy.

Claims

1. A system, comprising:

a first one or more computing nodes that, as result of processing a first update of a data item stored in a first collection of data items, transmits information indicative of the first update over a broadcast communications channel, wherein the information indicative of the first update is transmitted based at least in part on determining that the update originated from a client request to update the data item in the first collection of data items;

a second one or more computing nodes that, as a result of being activated, maintains a second collection of data configured as an at least partial replica of the first collection; and

one or more memories comprising computer-executable instructions that, as a result of execution, cause the system at least to;

receive the information indicative of the first update from the broadcast communications channel; and

transmit a request to update the data item in the second collection, the request comprising the information indicative of the first update and a sequence indicator, wherein the request comprises a flag indicative of the request being associated with replication of the updated data item, wherein the second one or more computing nodes determines to process the first update in response to a determination that the first update does not conflict with a prior update to the data item, the determination based at least in part on the sequence indicator being indicative of the first update occurring after the prior update, wherein the system determines to not transmit the request to update the data item in the second collection over the broadcast communications channel, based at least in part on the flag, in the request, that indicates the request is associated with replication of the data item.

2. The system of claim 1, further comprising one or more memories comprising computer-readable instructions that, as a result of execution by the one or more computing devices, cause the system at least to:

transmit a request for a subscription to information transmitted on the broadcast communications channel; and

receive the information indicative of the first update based at least in part on the subscription.

3. The system of claim 1, wherein the first one or more computing nodes, as a result of receiving a request to delete an additional data item, updates the additional data item to comprise information indicative of a deleted state.

4. The system of claim 1, wherein the request to update the data item in the second collection comprises the flag indicative of the request having been originated by the system.

5. The system of claim 4, further comprising one or more memories comprising computer-readable instructions that, as a result of execution by the one or more computing devices, cause the system at least to:

determine to not transmit the information indicative of the request to update the data item in the second collection on the broadcast communications channel based at least in part on the flag; and

determine to transmit information indicative of an additional request to update an additional item in the second collection based at least in part on determining that the flag is not set in the additional request.

6. A computer-implemented method, comprising:

receiving, from a broadcast communications channel, information indicative of a first update to a data item stored in a first collection of data items, wherein the information indicative of the first update was broadcast based at least in part on determining that the update originated from a client request to update the data item;

transmitting a request to update a version of the data item stored in a second collection of data items, wherein the request comprises a flag indicative of the request being associated with replication of the data item, wherein the version of the data item is updated based at least in part on a determination that the first update does not conflict with a prior update of the version of the data item in the second collection, wherein the determination is based at least in part on a conflict detection policy indicated by the request; and

determining to not transmit the received request over the broadcast communications channel, based at least in part on the flag, in the request, that indicates the request is associated with replication of the data item.

7. The computer-implemented method of claim 6, further comprising:

transmitting a request for a subscription to information transmitted on the broadcast communications channel; and

receiving the information indicative of the first update based at least in part on the subscription.

8. The computer-implemented method of claim 6, further comprising:

updating the first data item to comprise information indicative of a deleted state in response to a request to delete the data item.

9. The computer-implemented method of claim 6, wherein the request associated with replication is sent in response to another request initiated by a client.

10. The computer-implemented method of claim 6, further comprising:

receiving a request to replicate the first collection of items to a second region;

forming the second collection based at least in part on schema of the first collection;

performing a bulk copy of data items from the first collection of items to the second collection of data items;

storing information indicative of a time associated with a most recently updated data item included in the bulk copy; and

processing a plurality of requests to update versions of the data items stored in the second collection, the plurality of requests associated with updates to data items in the first collection processed after the time associated with a most recently updated data item included in the bulk copy.

11. The computer-implemented method of claim 10, wherein the bulk copy is performed in response to determining that a capacity of the broadcast communications channel would be exceeded.

12. The computer-implemented method of claim 6, further comprising:

delaying the receiving of the information indicative of a first update, wherein the delaying is based at least in part on supporting a rollback operation.

13. The computer-implemented method of claim 6, further comprising:

maintaining a coordinator process operating in a first region, the first region comprising one or more computing nodes maintaining the first collection; and

monitoring, by the coordinator process, updates to the first collection that have not been replicated to the second collection.

14. A non-transitory computer-readable storage medium having stored thereon instructions that, as a result of execution by one or more computing devices, cause the one or more computing devices at least to:

transmit a request to receive a stream of information indicative of updates to data items in a first collection of data items, the information to be transmitted over a broadcast communications channel;

receive, from the broadcast communications channel, information indicative of a first update to a data item stored in the first collection of data items;

determine that the first update was in response to a request to update the data item issued by a client; and

transmit a request to update a version of the data item stored in a second collection of data items, wherein the request comprises a flag indicative of the request being associated with replication, wherein the version of the data item is updated based at least in part on a determination that the first update does not conflict with a prior update of the version of the data item in the second collection, wherein the determination is based at least in part on a policy indicated by the request, and wherein it is determined to not transmit the request over the broadcast communications channel, based at least in part on the flag, in the request, that indicates the request is associated with replication.

15. The non-transitory computer-readable storage medium of claim 14, comprising further instructions that, as a result of execution by the one or more computing devices, cause the one or more computing devices to at least:

transmit a request for a subscription to information transmitted on the broadcast communications channel; and

receive the information indicative of the first update based at least in part on the subscription.

16. The non-transitory computer-readable storage medium of claim 14, comprising further instructions that, as a result of execution by the one or more computing devices, cause the one or more computing devices to at least:

update the first data item to comprise information indicative of a deleted state in response to a request to delete the data item.

17. The non-transitory computer-readable storage medium of claim 14, comprising further instructions that, as a result of execution by the one or more computing devices, cause the one or more computing devices to at least:

perform a bulk copy of data items from the first collection of data items to the second collection of data items;

store information indicative of a time associated with a most recently updated data item included in the bulk copy; and

process a plurality of requests to update versions of the data items stored in the second collection, the plurality of requests associated with updates to data items in the first collection of data items processed after the time associated with a most recently updated data item included in the bulk copy.

18. The non-transitory computer-readable storage medium of claim 14, comprising further instructions that, as a result of execution by the one or more computing devices, cause the one or more computing devices to at least:

maintain a coordinator process operating in a first region, the first region comprising one or more computing nodes maintaining the first collection of data items; and

monitor, by the coordinator process, updates to the first collection of data items that have not been replicated to the second collection of data items.

19. The non-transitory computer-readable storage medium of claim 14, wherein the policy comprises instructions that cause the version of the data item in the second collection to be updated when the first update occurred at a time later than a most recent update time of the version of the data item.

20. The non-transitory computer-readable storage medium of claim 14, wherein the version of the data item stored in the second collection is a transformed version of the data item.