Database remote replication for back-end tier of multi-tier computer systems

• US 7,370,064 B2
• Filed: 08/06/2002
• Issued: 05/06/2008
• Est. Priority Date: 08/06/2002
• Status: Expired due to Fees

• Current Assignee: Yousefi'Zadeh, Homayoun
• Original Assignee: Yousefi'Zadeh, Homayoun
• Inventors: Yousefi'zadeh, Homayoun
• Primary Examiner(s): Mofiz, Apu
• Assistant Examiner(s): CHEN, TE Y
• Application Number:
  US10/213,158
• Publication Number:
  US 20040030739A1
• Time in Patent Office: 2,100 Days
• Field of Search: 707/10, 707/102, 707/7, 709/217
• US Class Current: 1/1
• CPC Class Codes: G06F 11/202 where processing functionality is redundant redundant communication control functionality G06F11/2005, redundant storage control functionality G06F11/2089
• CPC Class Codes: G06F 11/2056 by mirroring
• CPC Class Codes: G06F 16/2471 Distributed queries

Abstract

A replication method and system for a computer system having multiple database servers for at least one database, wherein database servers are provided with a unified view of the data content. The method includes the steps of establishing connections to said multiple database servers for communicating with said database servers; distinguishing between read and update query requests; and assigning queries to respective ones of said multiple database servers to preserve the consistency of the respective unified view of data of said multiple database servers in case of receiving update queries. Assigning each new query includes the steps of distinguishing between read and update queries, determining possible assignments of that new query to one or more of said multiple database servers, each said possible update query assignment to one of said multiple database servers being based on a transaction-based replication scheme to preserve the unified view data of said multiple database servers; and assigning that new query to one of said multiple database servers as a function of said possible assignments of that new query, to preserve respective unified view of data of said multiple database servers.

Claims

1. A remote replication method for back-end tier of a multi-tier computer system comprising multiple database servers, wherein said replication method creates a unified view of data among all of said database servers in the back-end tier, and wherein said unified view provides a virtually centralized database server from a collection of physically distributed database servers by executing the following steps:

(i) designating at least one of said database servers as the primary replication server;

(ii) directly and constantly monitoring the health status of all of the existing database servers, detecting any non-functioning database server, and isolating said non-functioning database server from the rest of database servers;

(iii) receiving one or more database query requests (queries) from database clients including database functions for operating on data stored in said database servers;

(iv) distinguishing between read and update queries for said database query requests;

(v) establishing connections between said database servers and the primary replication server;

(vi) assigning each read query according to a load-balancing selection criterion to one of database servers for the purpose of communicating the read query results to the issuing database clients;

(vii) conditionally assigning each update query to said database servers comprising the substeps of;

a) placing transaction locks on the portion of data requested by the update query in all of said database servers,

b) synchronizing the update query among said database servers,

c) determining the synchronization status to release the transaction locks only when all of said database servers are synchronized with each other otherwise rolling back the changes from those database servers being updated, and

d) communicating the update query results to the issuing database clients.

2. The method of claim 1 further synchronizes the view of data of a newly added database server with the unified view of data of existing database servers prior to allowing said newly added database server function with said existing database servers online through a synchronization procedure which introduces no interruption of service in the operation of the back-end tier of said multi-tier computer system and comprises steps of:

(i) creating frequent synchronization checkpoints in said database servers;

(ii) maintaining a transient recovery change log to capture intermediate update query requests in all periods of the synchronization procedure processing;

(iii) synchronizing the newly added database server to the most recent synchronization checkpoint;

(iv) repeatedly analyzing and applying the remote replication method change log entries after synchronizing the newly added database server with the most recent synchronization checkpoint of said database servers; and

(v) finding a point in time containing no transaction in progress and bringing the newly added database server online.

3. The method of claim 1 wherein the logical step of establishing connections to said multiple database servers further comprises:

(i) using Java Database Connectivity (JDBC), Open Database Connectivity (ODBC) or a comparable database connectivity interface utilizing TCP/IP protocols for known IP addresses and port numbers;

(ii) maintaining said connections to said database servers; and

(iii) reusing existing connections for communication with said database servers.

4. The method of claim 3 wherein the logical step of establishing connections further includes the substeps of:

(i) establishing a first set of connections to a first subset of said multiple database servers using a first database connectivity interface; and

(ii) establishing a second set of connections to a second subset of said multiple database servers using a second database connectivity interface.

5. The method of claim 4 further comprising the steps of, after processing of a query request by a database server, receiving response from the database server in relation to said query request via the selected database sever, and communicating the response directly to the database client requester.

6. The method of claim 4, wherein establishing said connections comprises object oriented technologies with the steps of utilizing either a single-threaded or a multi-threaded inter-process communication protocol.

7. A remote database replication system with a remote replication method for back-end tier of a multi-tier computer system comprising multiple database servers and multiple remote replication servers wherein said replication system creates a unified view of data among all of said database servers in the back-end tier, and wherein said unified view provides a virtually centralized database server from a collection of physically distributed database servers by supplying means for:

(i) designating one of the database servers as the primary remote replication server;

(ii) monitoring the health status of the primary remote replication server by a secondary remote replication server and changing the designation of the primary remote replication server to said secondary replication server based on detecting a non-functioning status for said primary server;

(iii) receiving new database query requests from database clients including database functions for operating on data stored in said database servers;

(iv) distinguishing between read and update queries for said database query requests;

(v) establishing connections between said database servers and the primary replication server;

(vi) assigning each read query according to a load-balancing selection criterion to one of database servers for the purpose of communicating the read query results to the issuing database clients; and

(vii) conditionally assigning each update query to said database servers, placing transaction locks on the portion of data requested by the update query in all of said database servers, synchronizing the update query among said database servers, determining the synchronization status to release the transaction locks only when all of said database servers are synchronized with each other otherwise rolling back the changes from those database servers being updated, and communicating the update query results to the issuing database clients.

8. The remote replication system of claim 7 further includes means for synchronizing the view of data of a newly added database server with the unified view of data of existing database servers prior to allowing said newly added database server function with said existing database servers online through a synchronization procedure which introduces no interruption of service in the operation of the back-end tier of said multi-tier computer system and comprises steps of:

(i) creating frequent synchronization checkpoints in said database servers;

(ii) maintaining a transient recovery change log to capture intermediate update query requests in all periods of the synchronization procedure processing;

(iii) synchronizing the newly added database server to the most recent synchronization checkpoint;

(iv) repeatedly analyzing and applying the remote replication method change log entries after synchronizing the newly added database server with the most recent synchronization checkpoint of said database servers; and

(v) finding a point in time containing no transaction in progress and bringing the newly added database server online.

9. The remote replication system of claim 7 wherein means for establishing connections to said multiple database servers further comprises:

(i) utilizing Java Database Connectivity (JDBC), Open Database Connectivity (ODBC) or a comparable database connectivity interface utilizing TCP/IP protocols for known IP addresses and port numbers;

(ii) maintaining said connections to said database servers; and

(iii) reusing existing connections for communication with said database servers.

10. The remote replication system of claim 9 further comprises means for establishing connections which includes the substeps of:

(i) establishing a first set of connections to a first subset of said multiple database servers using a first database connectivity interface; and

(ii) establishing a second set of connections to a second subset of said multiple database servers using a second database connectivity interface.

11. The remote replication system of claim 10, wherein means for establishing said connections comprises object oriented technologies with the steps of utilizing either a single-threaded or a multi-threaded inter-process communication protocol.

12. The remote replication system of claim 7, wherein the multi-tier computer system includes multiple computers interconnected via a communication link means, at least one of said computers designated as a front-end tier computer generating said requests, at least two of said computers designated as a back-end tier computer, each back-end computer including a database server, and at least one of said computers designated as a middle tier computer for interfacing between the front-end and back-end tier computers.

13. The remote replication system of claim 12, wherein at least one of said multiple computers is designated as a remote replication server for interfacing between the middle tier and back-end tier computers.

14. The remote replication system of claim 12, wherein the set of single or multiple remote replication servers and multiple database servers further acts as a single database server to the middle tier servers and front-end computers.

15. The remote replication system of claim 12 further comprising means for, after processing of a query request by a database server, receiving response from the database server in relation to said query request via the selected database sever, and communicating the response indirectly to the requesting middle tier server via the remote replication server or directly to the requesting middle tier server.

16. A computer program product stored in a computer readable non-volatile and volatile storage medium for use with a database remote replication system in back-end tier of a multi-tier computer system comprising multiple database servers and multiple remote replication servers wherein said computer program creates a virtually centralized database server from a collection of physically distributed database servers and wherein said computer program comprises:

(i) means for designating one of the remote replication servers as the primary replication server;

(ii) means for monitoring the health status of all of the existing database servers directly by said primary replication server and isolating any non-functioning database server from the rest of database servers;

(iii) means for monitoring the health status of the primary remote replication server directly by a secondary replication server and replacing it with said secondary replication server in case of failure;

(iv) means for receiving one or more database query requests by the primary replication server requiring database functions for operating on data stored in said database servers;

(v) means for distinguishing between read and update queries for said database query requests;

(vi) means for utilizing either a single-threaded or a multi-threaded inter-process communication protocol connections from said primary remote replication server to said multiple database servers for communicating with said database servers;

(vii) means for assigning each read query according to a load-balancing selection criterion to one of database servers for the purpose of communicating the read query results to the issuing database clients;

(viii) means for conditionally assigning each update query to said database servers, placing transaction locks on the portion of data requested by the update query in all of said database servers, synchronizing the update query among said database servers, determining the synchronization status to release the transaction locks only when all of said database servers are synchronized with each other otherwise rolling back the changes from those database servers being updated, and communicating the update query results to the issuing database clients.

17. The computer program product of claim 16 further provides, means for synchronizing the view of data of a newly added database server with the unified view of data of existing database servers prior to allowing said newly added database server function with said existing database servers online through a synchronization procedure which introduces no interruption of service in the operation of the back-end tier of said multi-tier computer system and comprises steps of:

(i) creating frequent synchronization checkpoints in said database servers;

(ii) maintaining a transient recovery change log to capture intermediate update query requests in all periods of the synchronization procedure processing;

(iii) synchronizing the newly added database server to the most recent synchronization checkpoint;

(iv) repeatedly analyzing and applying the remote replication method change log entries after synchronizing the newly added database server with the most recent synchronization checkpoint of said database systems; and

(v) finding a point in time containing no transaction in progress and bringing the newly added database server online.

18. The computer program product of claim 16 wherein means for the logical step of establishing connections to said multiple database servers further comprises:

(i) using Java Database Connectivity (JDBC), Open Database Connectivity (ODBC) or a comparable database connectivity interface utilizing TCP/IP protocols for known IP addresses and port numbers;

(ii) maintaining said connections to said database servers; and

(iii) reusing existing connections for communication with said database servers.

19. The computer program product of claim 18 wherein the logical step of establishing connections further includes the substeps of:

(i) establishing a first set of connections to a first subset of said multiple database servers using a first database connectivity interface; and

(ii) establishing a second set of connections to a second subset of said multiple database servers using a second database connectivity interface.

20. The computer program product of claim 19 further comprising means for, after processing of query request by a database server, receiving response from the database server in relation to said query request via the selected database sever, and communicating the response indirectly to the requesting middle tier server via the remote replication server or directly to the requesting middle tier server.