Modeling distribution and failover database connectivity behavior

• US 8,996,909 B2
• Filed: 10/08/2009
• Issued: 03/31/2015
• Est. Priority Date: 10/08/2009
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Michailov, Zlatko V., Olix, David W., Neerincx, Matt A., Mumford, Chadwin J.
• Primary Examiner(s): Lottich, Joshua P
• Application Number:
  US12/575,475
• Publication Number:
  US 20110087636A1
• Time in Patent Office: 2,000 Days
• Field of Search: 709/223, 714/4.11, 714/4.12, 714/4.2, 714/6.22, 714/6.23, 714/6.3, 714/13
• US Class Current: 714/6.3
• CPC Class Codes: G06F 11/2023 Failover techniques
• CPC Class Codes: G06F 11/2094 Redundant storage or storage space G06F11/2056 takes precedence
• 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

Abstract

Aspects of the subject matter described herein relate to database connectivity behavior. In aspects, an entity seeking to access data that is replicated via instances of a database follows a plan. The plan provides information as to which instance(s) the entity is to attempt to access the data from. The plan may include connection data that indicates parameters the entity is to use in connecting to the instance(s). The plan may also indicate one or more other instances from which the entity may attempt to access the data if the client is not able to access the data from initial instance(s).

Claims

1. A method implemented on a client device having at least one processor, the method comprising:

seeking to access data of a database that is replicated by multiple instances of the database, the multiple instances hosted by multiple servers in multiple server farms, wherein the data is stored in at least one instance of a database that is accessed through a server in a server farm and which is replicated in another instance that is accessed through a different server in a different server farm;

accessing a plan to determine from which server the data can be accessed, the plan including distribution data, failover data, and connection data, the distribution data indicates which server to access the data, the failover data indicates which server to access the data when other instances are unavailable, the connection data indicates how to connect to a server hosting the data;

using the distribution data to determine that a first select server hosts the data;

attempting to access the first select server based on the connection data; and

upon failing to access the first select server, using the failover data to determine that a second select server hosts data.

2. The method of claim 1, wherein the failover data includes a failover process that proceeds to access the first select server for a first period of time and then proceeds to attempt to access the second select server for a second period of time when the first select server is unavailable.

3. The method of claim 1, wherein the failover data includes a failover process that proceeds to access both the first select server and the second select server simultaneously and to select whichever server responds first to access the data.

4. The method of claim 1, wherein the connection data includes a timeout value that indicates a length of time to wait to establish a connection.

5. The method of claim 1, wherein the distribution data includes at least one distribution process that is based on one of an ordered selection, a random selection or a round robin selection.

6. The method of claim 1, wherein the plan comprises a tree-structured representation having a plurality of nodes, a first subset of nodes including intermediate nodes, and a second subset of nodes including leaf nodes, each leaf node representing a server hosting an instance of a database and connection data, each intermediate node connected to one or more leaf nodes, each intermediate node representing a distribution process or a failover process associated with the leaf nodes connected thereto.

7. The method of claim 6, further comprising:

traversing the tree-structured representation to find a server to access the data.

8. A non-transitory computer-readable storage medium storing instructions that when executed on a processor causes the processor to perform actions comprising:

seeking to access data of a database that is replicated by multiple instances of the database, the multiple instances hosted by multiple servers in multiple server farms, wherein the data is stored in at least one instance that is accessed through a server in a server farm and which is replicated in another instance that is accessed through a different server in a different server farm;

accessing a plan to determine which server the data can be accessed, the plan including distribution data, failover data, and connection data, the distribution data indicates which server to access the data, the failover data indicates which server to access the data when other instances are unavailable, the connection data indicates how to connect to a server hosting the data;

using the distribution data to determine that a first select server hosts the data;

attempting to access the first select server based on the connection data; and

upon failing to access the first select server, using the failover data to determine that a second select server hosts the data.

9. The non-transitory computer-readable storage medium of claim 8, wherein the failover data includes a failover process that proceeds to access the first select server for a first period of time and then proceeds to attempt to access the second select server for a second period of time when the first select server is unavailable.

10. The non-transitory computer-readable storage medium of claim 8, wherein the failover data includes a failover process that proceeds to access both the first select server and the second select server simultaneously and to select whichever server responds first to access the requested data.

11. The non-transitory computer-readable storage medium device of claim 8, wherein the connection data includes a timeout value that indicates a length of time to wait to establish a connection.

12. The non-transitory computer-readable storage medium of claim 8, wherein the distribution data includes at least one distribution process that is based on one of an ordered selection, a random selection or a round robin selection.

13. The non-transitory computer-readable storage medium of claim 8, wherein the plan comprises a tree-structured representation having a plurality of nodes, a first subset of nodes including intermediate nodes, and a second subset of nodes including leaf nodes, each leaf node representing a server and connection data, each intermediate node connected to one or more leaf nodes, each intermediate node representing a distribution process or a failover process associated with the leaf nodes connected thereto.

14. The non-transitory computer-readable storage medium of claim 13, further comprising:

traversing the tree-structured representation to find the at least one server from which to access the data.

15. A system comprising:

at least one processor and at least one memory;

the at least one memory including the following processor-executable program modules;

a plan manager that obtains a plan to access data of a database that is replicated by multiple instances hosted by multiple servers in multiple server farms, wherein the data is stored in at least one instance that is accessed through a server in a server farm and which is replicated in another instance that is accessed through a different server in a different server farm, the plan including distribution data, failover data, and connection data, the distribution data indicates which server to access the data, the failover data indicates which server to access the data when other instances are unavailable, the connection data indicates how to connect to a server hosting the data;

a plan evaluator that utilizes the plan to determine a selected server from which to access the data;

a distribution manager that determines which distribution process to use to distribute a request to access the data;

a failover manager that determines from which server to access the data when the selected server is unavailable; and

a connection manager that accesses the data from the selected server using the connection data.

16. The system of claim 15, wherein the failover data includes a failover process that proceeds to access the selected server for a first period of time and then proceeds to attempt to access a second selected server for a second period of time when the first selected server is unavailable.

17. The system of claim 15, wherein the failover data includes a failover process that proceeds to access both the first selected server and the second selected server simultaneously and to select whichever server responds first to access the data.

18. The system of claim 15, wherein the plan comprises a tree-structured representation having a plurality of nodes, a first subset of nodes including intermediate nodes, and a second subset of nodes including leaf nodes, each leaf node representing a server and connection data, each intermediate node connected to one or more leaf nodes, each intermediate node representing a distribution process or a failover process associated with the leaf nodes connected thereto.

19. The system of claim 18, wherein the plan manager traverses the tree-structured representation to find a server from which to access the data.

20. The system of claim 18, wherein the plan manager traverses the tree-structured representation to find at least one failover server from which to access the data when the server is unavailable.