Method and system for transparent read-write query routing when load balancing databases

• US 9,032,017 B1
• Filed: 08/10/2010
• Issued: 05/12/2015
• Est. Priority Date: 08/10/2010
• Status: Active Grant

• Current Assignee: Ignite ScaleArc Solutions, Inc.
• Original Assignee: Scalarc Inc.
• Inventors: Singh, Varun, Sawant, Uday V., Goel, Prateek, Deshaveni, Naresh G.
• Primary Examiner(s): Pollack, Melvin H
• Application Number:
  US12/853,650
• Time in Patent Office: 1,736 Days
• Field of Search: 709/203
• US Class Current: 709/203
• CPC Class Codes: G06F 16/182 Distributed file systems
• CPC Class Codes: G06F 16/25 Integrating or interfacing systems involving database management systems
• CPC Class Codes: G06F 2209/5011 Pool
• CPC Class Codes: G06F 9/5055 considering software capabilities, i.e. software resources associated or available to the machine

Abstract

Roughly described, a database query router is installed transparently in a network between database client systems and a database server cluster. One subset of the servers is designated to handle read queries only and a second subset is designated to handle write queries. The query router receives, through a network port, the database queries made by a client system, and directed to the IP address and port assigned to the query router. The query router determines whether the query is a read or a write query, and forwards read queries to servers in the first subset and write queries to servers in the second subset of servers. By installing the query router transparently in the network, no modifications are required to the software in either the client systems or the database servers in order to add query routing functionality to a client/server database environment.

Claims

1. A database accelerator, for use in a network having a set of at least one client system each running an application which makes database queries via a network to a particular IP address and port combination, and a plurality of database servers each having access to at least a portion of a database, the database accelerator being arranged to:

designate a first subset of at least one of the database servers to handle read queries only, and a second subset of at least one of the database servers to handle write queries;

receive, through a network port, the database queries made by the set of client systems and directed to the particular IP address and port combination;

determine whether each of a plurality of the received queries is a read query or a write query;

forward to one or more of the servers in the second subset of servers each of the received queries in the plurality of the received queries which are determined to be write queries, and

forward to one or more servers in the first subset of servers at least one of the received queries in the plurality of the received queries which are determined to be read queries.

2. A database accelerator according to claim 1, wherein a particular server in the second subset of servers is further designated to handle read queries as well,

and wherein the database accelerator is further arranged to forward to one or more servers in the second subset of servers at least one of the received queries in the plurality of the received queries which are determined to be read queries.

3. A database accelerator according to claim 1, wherein the allocation of the database servers to the first or second subset is made in response to user input.

4. A database accelerator according to claim 1, wherein the plurality of database servers is organized in master/slave configuration,

and wherein all of the database servers in the plurality which are masters are allocated to the second subset,

and wherein all of the database servers in the plurality which are slaves are allocated to the first subset.

5. A database accelerator according to claim 1, wherein each of the servers in the plurality of database servers has a respective IP address and port combination at which it receives database queries,

and wherein the database accelerator is arranged to forward queries to a given server in the plurality of database servers by forwarding them to the IP address and port combination at which the given server receives database queries.

6. A database accelerator according to claim 1, embodied in a physical enclosure separate from the client systems and the database servers.

7. A database accelerator according to claim 1, wherein the network port through which the database accelerator receives the database queries made by the set of client systems and directed to the particular IP address and port combination, is a physical network port.

8. A database accelerator according to claim 1, comprising a network port through which the database accelerator forwards received queries to servers in the first and second sets of servers.

9. A database accelerator according to claim 8, wherein the network port through which the query router forwards received queries to servers in the first and second sets of servers, is a physical network port.

10. A database accelerator according to claim 1, embodied in a virtual machine.

11. A database accelerator according to claim 10, wherein the network port through which the query router receives the database queries made by the set of client systems and directed to the particular IP address and port combination, is a virtual network port.

12. A database accelerator according to claim 10, wherein the virtual machine is housed in a common enclosure with one of the client systems.

13. A database accelerator according to claim 10, wherein the virtual machine is housed in a common enclosure with one of the plurality of database servers.

14. A database accelerator according to claim 1, wherein:

either one of the client systems is embodied in a virtual machine and housed in a common enclosure with another of the client systems;

or one of the client systems is embodied in a virtual machine and housed in a common enclosure with one of the plurality of database servers;

or one of the plurality of database servers is embodied in a virtual machine and housed in a common enclosure with another one of the plurality of database servers.

15. A database accelerator according to claim 1, further arranged to receive responses from the plurality of database servers to queries forwarded by the database accelerator;

and to forward the received responses to originating ones of the set of client systems.

16. A database accelerator according to claim 1, wherein a first one of the plurality of database servers comprises:

a first network port at which the first server receives all database queries directed to the first server; and

a plurality of subsystems each having a respective different IP address and port combination, and throughout which the first server distributes workload for executing database queries received by the first server.

17. A query routing method, for use in a network having a set of at least one client system each running an application which makes database queries via a network to a particular IP address and port combination, and a plurality of database servers each having access to at least a portion of a database, comprising the steps of:

designating a first subset of at least one of the database servers to handle read queries only and a second subset of at least one of the database servers to handle write queries;

a database accelerator receiving, through a network port, the database queries made by the set of client systems and directed to the particular IP address and port combination;

the database accelerator determining whether each of a plurality of the received queries are read queries or write queries;

the database accelerator forwarding to one or more of the servers in the second subset of servers each of the received queries in the plurality of the received queries which are determined to be write queries, and

the database accelerator forwarding to one or more servers in the first subset of servers at least one of the received queries in the plurality of the received queries which are determined to be read queries.

18. A method according to claim 17, wherein a particular server in the second subset of servers is further designated to handle read queries as well,

further comprising the step of the database accelerator forwarding to one or more servers in the second subset of servers at least one of the received queries in the plurality of the received queries which are determined to be read queries.

19. A method according to claim 17, further comprising the step of allocating the database servers to the first or second subset in dependence upon user input.

20. A method according to claim 17, wherein the plurality of database servers is organized in master/slave configuration, further comprising the steps of:

allocating to the second subset all of the database servers in the plurality which are masters,

and allocating to the first subset all of the database servers in the plurality which are slaves.

21. A method according to claim 17, wherein each of the servers in the plurality of database servers has a respective IP address and port combination at which it receives database queries,

and wherein the steps of forwarding to one or more of the servers, each comprises the step of forwarding to the IP address and port combination at which each destination server receives database queries.

22. A method according to claim 17, wherein the database accelerator is embodied in a physical enclosure separate from the client systems and the database servers.

23. A method according to claim 17, wherein the network port through which the database accelerator receives database queries in the step of the database accelerator receiving the database queries, comprises a physical network port.

24. A method according to claim 17, wherein the step of the database accelerator forwarding to a server in the second subset of servers each of the received queries determined to be write queries, comprises the step of the database accelerator forwarding each of such received queries through a physical network port.

25. A method according to claim 17, further comprising the steps of:

the database accelerator receiving responses from the plurality of database servers to queries forwarded by the database accelerator; and

the database accelerator forwarding the received responses to originating ones of the set of client systems.

26. A computer readable medium having stored thereon in a non-transitory manner, a plurality of software code portions defining logic for performing the method of claim 17.

27. A computer readable medium having stored thereon in a non-transitory manner, a plurality of software code portions defining logic for performing the method of claim 19.

28. A computer readable medium having stored thereon in a non-transitory manner, a plurality of software code portions defining logic for performing the method of claim 21.

29. A computer readable medium having stored thereon in a non-transitory manner, a plurality of software code portions defining logic for performing the method of claim 25.

30. A database accelerator, for use in a network having a set of at least one client system each running an application which makes database queries via a network to a particular IP address and port combination, and a plurality of database servers each having access to at least a portion of a database, the database accelerator further comprising:

means to designate a first subset of at least one of the database servers to handle read queries only, and a second subset of at least one of the database servers to handle write queries;

means to receive, through a network port, the database queries made by the set of client systems and directed to the particular IP address and port combination;

means to determine whether each of a plurality of the received queries is a read query or a write query;

means to forward to one or more of the servers in the second subset of servers each of the received queries in the plurality of the received queries which are determined to be write queries, and

means to forward to one or more servers in the first subset of servers at least one of the received queries in the plurality of the received queries which are determined to be read queries.