APPARATUS AND METHODS OF A ZERO SINGLE POINT OF FAILURE LOAD BALANCER

• US 20080133687A1
• Filed: 11/28/2007
• Published: 06/05/2008
• Est. Priority Date: 12/05/2006
• Status: Active Grant

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Yip, Eric Chi Chung, Fok, Kenny
• Granted Patent: US 8,732,261 B2
• Time in Patent Office: Days
• US Class Current: 709/207
• CPC Class Codes: G06F 11/2023 Failover techniques
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals
• CPC Class Codes: H04L 12/40019 Details regarding a bus master
• CPC Class Codes: H04L 41/0668 by dynamic selection of recovery network elements, e.g. replacement by the most appropriate element after failure
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1031 Controlling of the operation of servers by a load balancer, e.g. adding or removing servers that serve requests
• CPC Class Codes: H04L 67/1034 Reaction to server failures by a load balancer
• CPC Class Codes: H04L 67/59 Providing operational support to end devices by off-loading in the network or by emulation, e.g. when they are unavailable
• CPC Class Codes: H04L 69/40 for recovering from a failure of a protocol instance or entity, e.g. service redundancy protocols, protocol state redundancy or protocol service redirection management of faults, events, alarms or notifications in data switching networks H04L41/06

Abstract

Apparatus and methods in a load balancer computing apparatus include, in aspects, determining if a master is functioning, and if the master is not functioning, then attempting to become the master by setting a master network identification in a master database table to a self network identification before at least one other slave sets a corresponding other slave network identification as the master network identification in the master database table. In other aspects, the apparatus and methods include assigning each of a group of commands to one of a plurality of slaves, storing the commands and a corresponding assignment in a database table, checking a command status in the database table for each of the group of commands, wherein the command status indicates whether or not the respective command has been executed; and executing a callback function when all the commands are executed.

Claims

1. A computer-implemented method for a slave computing apparatus of a load balancer, comprising:

determining, according to a common predetermined health maintenance algorithm independently operable by each of a plurality of slaves in the load balancer, if a master of the load balancer is functioning; and

if the master is not functioning, then attempting to become the master according to a common predetermined role transitioning algorithm independently operable by each of the plurality of slaves in the load balancer, wherein the predetermined role transitioning algorithm includes setting a master network identification in a master database table to a self network identification before at least one other slave sets a corresponding other slave network identification as the master network identification in the master database table.

2. The method of claim 1, wherein attempting to become the master further includes setting the self network identification in a portion of the master database table having a master code, wherein the master code comprises a predetermined value for a primary key of the master database table.

3. The method of claim 1, wherein attempting to become the master further comprises implementing a semaphore by using a portion of the master database table to control access to the master network identification.

4. The method of claim 1, wherein attempting to become the master further comprises:

reading the master network identification from a portion of the master database table based on a master code; and

attempting to delete the portion of the master database table comprising the master code and the master network identification.

5. The method of claim 1, further comprising:

reading the master network identification from a portion of the master database table corresponding to a master code;

checking if the master network identification corresponds to the self network identification;

if the master network identification does not correspond to the self network identification, then working as a slave; and

else if the master network identification corresponds to the self network identification, then working as the master.

6. The method of claim 1, where determining if the master is functioning, further comprises:

sending a communication to the master based on the master network identification; and

if an acknowledgement is not received from the master in a predetermined time, then attempting to become the master.

7. The method of claim 1, wherein determining if the master is functioning, further comprises sending a communication message to the master using transport control protocol/internet protocol (TCP/IP) over a local area network (LAN).

8. The method of claim 1, further comprising:

receiving a communication message from the master; and

sending an acknowledgement to the master.

9. The method of claim 1, further comprising:

receiving from the master at least one command to execute; and

sending a result of executing the command to the master.

10. The method of claim 9, further comprising executing the command, wherein executing the command comprises processing a performance statistic corresponding to at least one wireless device operating on a wireless network.

11. The method of claim 1, further comprising:

obtaining a master performance requirement; and

obtaining an actual performance measurement of the slave; and

allowing the slave to become the master only if the actual performance measurement meets the master performance requirement.

22. The slave apparatus of claim 1, further comprising:

receiving a communication message from the master; and

sending an acknowledgement to the master.

12. At least one processor configured to act as a slave of a load balancer, comprising:

a first module for determining, according to a common predetermined health maintenance algorithm independently operable by each of a plurality of slaves in the load balancer, if a master of the load balancer is functioning; and

a second module for attempting, if the master is not functioning, to become the master according to a common predetermined role transitioning algorithm independently operable by each of the plurality of slaves in the load balancer, wherein the predetermined role transitioning algorithm includes setting a master network identification in a master database table to a self network identification before at least one other slave sets a corresponding other slave network identification as the master network identification in the master database table.

13. A computer program product for a slave of a load balancer, comprising:

a computer-readable medium comprising;

a first set of codes for causing a computer to determine, according to a common predetermined health maintenance algorithm independently operable by each of a plurality of slaves in the load balancer, if a master of the load balancer is functioning; and

a second set of codes for causing the computer to attempt, if the master is not functioning, to become the master according to a common predetermined role transitioning algorithm independently operable by each of the plurality of slaves in the load balancer, wherein the predetermined role transitioning algorithm includes setting a master network identification in a master database table to a self network identification before at least one other slave sets a corresponding other slave network identification as the master network identification in the master database table.

14. A slave apparatus of a load balancer, comprising:

means for determining, according to a common predetermined health maintenance algorithm independently operable by each of a plurality of slaves in the load balancer, if a master of the load balancer is functioning; and

means for attempting, if the master is not functioning, to become the master according to a common predetermined role transitioning algorithm independently operable by each of the plurality of slaves in the load balancer, wherein the predetermined role transitioning algorithm includes setting a master network identification in a master database table to a self network identification before at least one other slave sets a corresponding other slave network identification as the master network identification in the master database table.

15. A slave apparatus of a load balancer, comprising:

a health management service operable to determine, according to a common predetermined health maintenance algorithm independently operable by each of a plurality of slaves in the load balancer if a master is functioning and further operable, if the master is not functioning, to attempt to become the master according to a common predetermined role transitioning algorithm independently operable by each of the plurality of slaves in the load balancer wherein the predetermined role transitioning algorithm includes setting a master network identification in a master database table to a self network identification before at least one other slave sets a corresponding other slave network identification as the master network identification in the master database table; and

a database service in communication with the health management service and operable to read and write to the master database table responsive to the health management service.

16. The slave apparatus of claim 15, wherein in response to the health management service, the database service is further operable to set the self network identification in a portion the master database table having a master code, wherein the master code comprises a predetermined value for a primary key of the master database table.

17. The slave apparatus of claim 15, wherein in response to the health management service, the database service is further operable to implement a semaphore by using a portion of the master database table to control access to the master network identification.

18. The slave apparatus of claim 15, wherein in response to the health management service, the database service is further operable to:

read the master network identification from a portion of the master database table based on a master code; and

attempt to delete the portion of the master database table comprising the master code and the master network identification.

19. The slave apparatus of claim 15, wherein at least one of the health management service or the database service is further operable to:

read the master network identification from a portion of the master database table corresponding to a master code;

check if the master network identification corresponds to the self network identification;

if the master network identification does not correspond to the self network identification, then cause the slave apparatus to work as a slave; and

else if the master network identification corresponds to the self network identification, then cause the slave apparatus to work as the master.

20. The slave apparatus of claim 15, wherein the health management service is further operable to:

send a communication to the master based on the master network identification; and

if an acknowledgement is not received from the master in a predetermined time, then attempt to become the master.

21. The slave apparatus of claim 15, wherein the health management service is further operable to send a communication message to the master using transport control protocol/internet protocol (TCP/IP) over a local area network (LAN).

23. The slave apparatus of claim 15, further comprising:

a user command executor operable to receive a command from the master over a communication network, execute the command, and send a result of executing the command to the master over the communications network.

24. The slave apparatus of claim 15, wherein the user command executor is further operable to process a performance statistic corresponding to at least one wireless device operating on a wireless network.

25. The slave apparatus of claim 15, further comprising a memory having a master performance requirement and an actual performance measurement of the slave, wherein the health maintenance service is further operable to allow the slave to become the master if the actual performance measurement meets the master performance requirement.

26. A computer-implemented method for a master computing apparatus of a load balancer, comprising:

executing a common predetermined role transitioning algorithm to become a master, wherein the common predetermined role transitioning algorithm is independently operable by each of a plurality of slaves in the load balancer and includes setting a master network identification in a master database table to a self network identification before at least one other slaves sets a corresponding other slave network identification as the master network identification in the master database table;

executing a common predetermined health maintenance algorithm to determine health of each of the plurality of slaves, wherein the common predetermined health maintenance algorithm is independently operable by each of the plurality of slaves in the load balancer;

assigning each of a group of commands to at least one of the plurality of slaves determined to be functioning based on the execution of the predetermined health maintenance algorithm;

storing the commands and a corresponding assignment in a database table, wherein each assignment comprises a network identification for one of the plurality of slaves assigned to the respective command;

checking a command status in the database table for each of the group of commands, wherein the command status indicates whether or not the respective command has been executed; and

executing a callback function when all the commands are executed as determined based on the command status.

27. The method of claim 26, wherein executing the callback function further comprises executing a command script on at least one result generated by execution of each of the group of commands.

28. The method of claim 26, wherein executing the common predetermined health maintenance algorithm further comprises:

determining if each of the plurality of slaves is functioning; and

if a respective one of the plurality of slaves is not functioning, then re-assigning each respective command assigned to the respective slave not functioning to at least one of the plurality of slaves that is functioning, wherein executing the callback function is based on the re-assigning.

29. The method of claim 28, wherein determining if each of the slaves is functioning, further comprises:

sending a communication message to each of the plurality of slaves; and

if an acknowledgement is not received from each of the plurality of slaves in a predetermined time, setting the slave as not functioning.

30. The method of claim 29, wherein sending further comprises sending a communication message to each of the plurality of slaves using transport control protocol/internet protocol (TCP/IP) over a local area network (LAN).

31. The method of claim 26, wherein assigning further comprises assigning based on a number of commands already assigned to each of the plurality of slaves.

32. The method of claim 26, wherein assigning further comprises assigning based on a utilization of each of the plurality of slaves.

33. The method of claim 26, further comprising building a slave identification table by using received communication messages from each of the plurality of slaves, wherein the slave identification table includes a network identification for each slave.

34. The method of claim 26, further comprising executing the common predetermined health maintenance algorithm prior to becoming the master to determine a health of an existing master, and wherein the executing of the common predetermined role transitioning algorithm to become the master is triggered based on a determination that the existing master is not functioning.

35. The method of claim 34, wherein becoming the master further comprises:

determining an actual performance measure;

determining a master performance requirement; and

if the actual performance measure meets the master performance requirement, then acting as the master.

36. At least one processor configured to act as a master of a load balancer comprising:

a first module for executing a common predetermined role transitioning algorithm to become a master, wherein the common predetermined role transitioning algorithm is independently operable by each of a plurality of slaves in the load balancer and includes setting a master network identification in a master database table to a self network identification before at least one other slaves sets a corresponding other slave network identification as the master network identification in the master database table;

a second module for executing a common predetermined health maintenance algorithm to determine health of each of the plurality of slaves, wherein the common predetermined health maintenance algorithm is independently operable by each of the plurality of slaves in the load balancer;

a third module for assigning each of a group of commands to at least one of the plurality of slaves determined to be functioning based on the execution of the predetermined health maintenance algorithm;

a fourth module for storing the commands and a corresponding assignment in a database table, wherein each assignment comprises a network identification for one of the plurality of slaves assigned to the respective command;

a fifth module for checking a command status in the database table for each of the group of commands, wherein the command status indicates whether or not the respective command has been executed; and

a sixth module for executing a callback function when all the commands are executed as determined based on the command status.

37. A computer program product, comprising:

a computer-readable medium comprising;

a first set of codes causing a computer to execute a common predetermined role transitioning algorithm to become a master, wherein the common predetermined role transitioning algorithm is independently operable by each of a plurality of slaves in the load balancer and includes setting a master network identification in a master database table to a self network identification before at least one other slaves sets a corresponding other slave network identification as the master network identification in the master database table;

a second set of codes causing the computer to execute a common predetermined health maintenance algorithm to determine health of each of the plurality of slaves, wherein the common predetermined health maintenance algorithm is independently operable by each of the plurality of slaves in the load balancer;

a third set of codes causing the computer to assign each of a group of commands to at least one of the plurality of slaves determined to be functioning based on the execution of the predetermined health maintenance algorithm;

a fourth set of codes causing the computer to store the commands and a corresponding assignment in a database table, wherein each assignment comprises a network identification for one of the plurality of slaves assigned to the respective command;

a fifth set of codes causing the computer to check a command status in the database table for each of the group of commands, wherein the command status indicates whether or not the respective command has been executed; and

a sixth set of codes causing the computer to execute a callback function when all the commands are executed as determined based on the command status.

38. A master apparatus of a load balancer comprising:

means for executing a common predetermined role transitioning algorithm to become a master, wherein the common predetermined role transitioning algorithm is independently operable by each of a plurality of slaves in the load balancer and includes setting a master network identification in a master database table to a self network identification before at least one other slaves sets a corresponding other slave network identification as the master network identification in the master database table;

means for executing a common predetermined health maintenance algorithm to determine health of each of the plurality of slaves, wherein the common predetermined health maintenance algorithm is independently operable by each of the plurality of slaves in the load balancer;

means for assigning each of a group of commands to at least one of the plurality of slaves determined to be functioning based on the execution of the predetermined health maintenance algorithm;

means for storing the commands and a corresponding assignment in a database table, wherein each assignment comprises a network identification for one of the plurality of slaves assigned to the respective command;

means for checking a command status in the database table for each of the group of commands, wherein the command status indicates whether or not the respective command has been executed; and

means for executing a callback function when all the commands are executed as determined based on the command status.

39. A master apparatus of a load balancer comprising:

a health maintenance service operable to execute a common predetermined role transitioning algorithm to become a master, wherein the common predetermined role transitioning algorithm is independently operable by each of a plurality of slaves in the load balancer and includes setting a master network identification in a master database table to a self network identification before at least one other slaves sets a corresponding other slave network identification as the master network identification in the master database table;

wherein the health maintenance service is further operable to execute a common predetermined health maintenance algorithm to determine health of each of the plurality of slaves, wherein the common predetermined health maintenance algorithm is independently operable by each of the plurality of slaves in the load balancer;

a distribution service operable to assign each of a group of commands to at least one of the plurality of slaves determined to be functioning based on the execution of the predetermined health maintenance algorithm;

a database service in communication with the distribution service and operable to store the commands and a corresponding assignment in a database table, wherein each assignment comprises a network identification for one of the plurality of slaves assigned to the respective command;

wherein the distribution service is further operable to check a command status in the database table for each of the group of commands, wherein the command status indicates whether or not the respective command has been executed; and

a user command executor operable to execute a callback function when all the commands are executed as determined based on the command status.

40. The master apparatus of claim 39, wherein the callback function comprises a command script executable on at least one result generated by execution of each of the group of commands.

41. The master apparatus of claim 39, further comprising a health maintenance service in communication with the distribution service, wherein the health maintenance service is operable to:

determine if each of the plurality of slaves is functioning; and

if a respective one of the plurality of slaves is not functioning, then to re-assign each respective command assigned to the respective slave not functioning to at least one of the plurality of slaves that is functioning, wherein execution of the callback function is based on the re-assignment.

42. The master apparatus of claim 41, wherein the health maintenance service is further operable to:

send a communication message to each of the plurality of slaves; and

if an acknowledgement is not received from each of the plurality of slaves in a predetermined time, then set the slave as not functioning.

43. The master apparatus of claim 42, wherein the health maintenance service is further operable to send a communication message to each of the plurality of slaves using transport control protocol/internet protocol (TCP/IP) over a local area network (LAN).

44. The master apparatus of claim 39, wherein the distribution service is further operable to assign based on a number of commands already assigned to each of the plurality of slaves.

45. The master apparatus of claim 39, wherein the distribution service is further operable to assign based on a utilization of each of the plurality of slaves.

46. The master apparatus of claim 39, wherein the database service is further operable to build a slave identification table by using received communication messages from each of the plurality of slaves, wherein the slave identification table includes a network identification for each slave.

47. The master apparatus of claim 39, wherein, prior to becoming the master, the health maintenance service is further operable to execute the common predetermined health maintenance algorithm to determine a health of an existing master, and wherein the health maintenance service is triggered to execute the common predetermined role transitioning algorithm to become the master based on a determination that the existing master is not functioning.

48. The master apparatus of claim 47, further comprising a memory having an actual performance measure and a master performance requirement, wherein the health maintenance service is further operable to enable the master apparatus to act as the master if the actual performance measure meets the master performance requirement.