System and method of distributing replication commands

• US 20040162859A1
• Filed: 02/19/2003
• Published: 08/19/2004
• Est. Priority Date: 02/19/2003
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Guo, Qun, Habben, Michael E.
• Granted Patent: US 7,707,181 B2
• Time in Patent Office: Days
• US Class Current: 707/204
• 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: G06F 16/275 Synchronous replication

Abstract

A transactional replication system wherein each replicated command applies to a single row as identified by the primary key, replicated commands are dispatched among multiple connections by hashing of primary key from a single command queue to ensure that changes to the same data row are always dispatched to the same connection to maintain the order of commands as originally occurred on publisher. All connections commit their transactions at the same time in a coordinated fashion without a two phase commit. Retry logic based on a logical sequence number is used to reapply commands from any connection that fails to commit.

Claims

1. A system for distributing replication commands, comprising:

a distributor having a command queue that receives replication commands from a publisher, the distributor further having multiple connections to communicate the replication commands to subscribers; and

a plurality of command buckets, each of the command buckets communicating to the distributor via one of the multiple connections, each of the plurality of command buckets receiving only those replication commands to be applied to a unique row of a table of the subscribers, the received replication commands each containing an identifier of the unique row.

2. The system of claim 1, wherein the publisher decorates each of the replication commands with a hash of the primary key of the row of a table to be operated upon.

3. The system of claim 2, wherein each of the replication commands are sorted into a respective command bucket in accordance with the hash of the primary key by a dispatcher thread.

4. The system of claim 2, wherein the distributor marks each bucket ready for a commit, and wherein the replication commands in each bucket are applied to the subscriber at substantially the same time.

5. The system of claim 4, wherein an executor of each bucket notifies the distributor that the commit has been completed such that the distributor resumes processing.

6. The system of claim 4, further comprising a retry mechanism to determine if each of the connections has successfully committed.

7. The system of claim 6, wherein if a connection has failed to commit, the retry mechanism restarts the commit process from an earliest failed point in accordance with a logical sequence number of the replication commands, and applies the replication commands having newer logical sequence numbers than a last logical sequence number of a replication command to successfully commit.

8. A method for distributing replication commands in a system having a subscriber, distributor and publisher, comprising:

decorating replication commands at the publisher with a unique identifier of a row on which the replication commands are to be operated upon;

forwarding the replication commands to the distributor;

storing the replication commands in a command queue; and

forwarding the replication commands to a plurality of command buckets, each of the plurality of command buckets having a connection to the distributor, wherein each command bucket receives replication commands with the same hash key

9. The method of claim 8, wherein the hash key is a hash of the primary key of the row of a table to be operated upon.

10. The method of claim 9, wherein forward the replication commands to the plurality of command buckets further comprises sorting each of the commands into a respective command bucket in accordance with the hash of the primary key.

11. The method of claim 9, further comprising:

marking each command bucket ready for a commit; and

applying the replication commands in each bucket to the subscriber at substantially the same time.

12. The method of claim 11, further comprising:

notifying the distributor that each bucket has completed a commit process; and

resuming processing at the distributor.

13. The method of claim 11, further comprising:

determining if the replication commands of each bucket have committed successfully at the subscriber.

14. The method of claim 13, further comprising:

restarting the commit process from an earliest failed point in accordance with a logical sequence number of the replication commands; and

applying the replication commands having newer logical sequence numbers than a last logical sequence number of a replication command to successfully commit.

15. A computer-readable medium having computer-executable instructions for performing a method for distributing replication commands in a system having a subscriber, distributor and publisher, comprising:

decorating replication commands at the publisher with a hash key which is a hashed value of primary key, the primary key uniquely identifies a row on which the replication commands are to be operated upon.

forwarding the replication commands to the distributor;

storing the replication commands in a command queue;

forwarding the replication commands to a plurality of command buckets, each of the plurality of command buckets having a connection to the distributor, wherein each command bucket receives replication commands with the same hash key.

16. The computer-readable medium of claim 15, wherein the hash key is a hash of the primary key of the row of a table to be operated upon.

17. The computer-readable medium of claim 16, wherein forward the replication commands to the plurality of command buckets further comprises sorting each of the commands into a respective command bucket in accordance with the hash of the primary key.

18. The computer-readable medium of claim 16, further comprising instructions for:

marking each command bucket ready for a commit; and

applying the replication commands in each bucket to the subscriber at substantially the same time.

19. The computer-readable medium of claim 18, further comprising instructions for:

notifying the distributor that each bucket has completed a commit process; and

resuming processing at the distributor.

20. The computer-readable medium of in claim 18, further comprising instructions for:

determining if the replication commands of each bucket have committed successfully at the subscriber.

21. The computer-readable medium of claim 20, further comprising instructions for:

restarting the commit process from an earliest failed point in accordance with a logical sequence number of the replication commands; and

applying the replication commands having newer logical sequence numbers than a last logical sequence number of a replication command to successfully commit.