Dynamic monitor and controller of availability of a load-balancing cluster

• US 7,296,268 B2
• Filed: 12/18/2000
• Issued: 11/13/2007
• Est. Priority Date: 12/18/2000
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Darling, Christopher L., Gernaey, Michael E., Aikins, Howard A., Kaldestad, Hallvard S.
• Primary Examiner(s): Thomson; William
• Assistant Examiner(s): Ho; Andy
• Application Number:
  US09/740,720
• Publication Number:
  US 20020078263A1
• Time in Patent Office: 2,521 Days
• Field of Search: 718/105, 718/106, 709/230
• US Class Current: 718/105
• CPC Class Codes: G06F 2209/505 Clust
• CPC Class Codes: G06F 9/5061 Partitioning or combining of resources

Abstract

An implementation of a technology, described herein, for remotely and dynamically monitoring the availability of the members of a load-balancing cluster. An implementation of the claimed invention includes a dynamic, exocluster application-layer monitor for dynamically monitoring and dynamically controlling the members of a load-balancing cluster. It may actively control the members that it monitors. The exocluster monitor is protocol agnostic. It dynamically adjusts so that it can monitor all of the members of the cluster as members are added and removed. This abstract itself is not intended to limit the scope of this patent. The scope of the present invention is pointed out in the appending claims.

Claims

1. A method performed by a control client that communicates with a load-balancing cluster of server nodes configured to service application-layer requests sent by user clients to a virtual address common to the cluster, the method comprising:

dynamically determining, by the control client, which server nodes are members of the load-balancing cluster;

monitoring application-layer availability of the server nodes in the load-balancing cluster, the monitoring being performed by the control client generating and sending, from outside of the load-balancing cluster application-layer, requests to those server nodes determined to be members of the load-balancing cluster; and

in accordance with the application-layer availability of a server node in the load-balancing cluster as determined by the application-layer monitoring of the control client, sending to the load-balancing cluster a message that is configured to control whether the server node handles application-layer requests sent to the load-balancing cluster by the user clients, where the application-layer requests sent by the control client to monitor application-layer availability conform to a same application-layer protocol that the user client requests conform to.

2. A method as recited in claim 1 further comprising exoclusterly and selectively deactivating one or more active members of the cluster.

3. A method as recited in claim 1 further comprising, based upon the monitoring, identifying one or more active members of the cluster that are presently overwhelmed at the application-layer.

4. A method as recited in claim 1 further comprising:

based upon the monitoring, identifying one or more active members of the cluster that are presently overwhelmed at the application-layer;

exoclusterly deactivating one or more members identified by the identifying.

5. A method as recited in claim 1 further comprising exoclusterly and selectively activating one or more inactive members of the cluster.

6. A method as recited in claim 1 further comprising, based upon the monitoring, identifying one or more inactive members of the cluster that are not presently overwhelmed at the application-layer.

7. A method as recited in claim 1 further comprising:

based upon the monitoring, identifying one or more inactive members of the cluster that are not presently overwhelmed at the application-layer;

exoclusterly activating one or more members identified by the identifying.

8. A method as recited in claim 1 further comprising:

based upon the monitoring, identifying one or more active members of the cluster that are presently overwhelmed at the application-layer and identifying one or more inactive members of the cluster that are not presently overwhelmed at the application-layer;

exoclusterly deactivating one or more active members identified by the identifying;

exoclusterly activating one or more inactive members identified by the identifying.

9. A method as recited in claim 1 further comprising determining a present activity state of members of the cluster.

10. A method as recited in claim 1 further comprising:

determining a present activity state of members of the cluster;

tracking and persisting the activity states of the members of the cluster.

11. A method as recited in claim 10, wherein the activity states include cluster states.

12. A method as recited in claim 10 further comprising reporting a present activity state of one or more members of the cluster.

13. A method as recited in claim 10 further comprising reporting historical record of the activity states of one or more members of the cluster.

14. A method as recited in claim 10 further comprising reporting a present application-layer state of one or more members of the cluster.

15. A method as recited in claim 10 further comprising reporting historical record of the application-layer states of one or more members of the cluster.

16. A method as recited in claim 1, wherein the monitoring comprises monitoring in one or more different application-layer protocols.

17. A method as recited in claim 1, further comprises, based upon the monitoring, determining the application-layer availability of one or more members based upon an indicator of such availability, the indicator sent from a member being monitored.

18. A method as recited in claim 1, further comprises:

based upon the monitoring, determining the application-layer availability of one or more members based upon a indicator of such availability, the indicator sent from a member being monitored;

the member being monitored determining such availability and generating such indicator.

19. A computer-readable storage medium having computer-executable instructions that, when executed by a computer, perform the method as recited in claim 1.

20. A method performed by a control client that communicates with a load-balancing cluster of server nodes configured to service application-layer requests sent by user clients to a virtual address common to the cluster, the method comprising:

monitoring application-layer availability of the server nodes in the load-balancing cluster, the monitoring being performed by the control client generating and sending, from outside of the load-balancing cluster application-layer, requests to those server nodes determined to be members of the load-balancing cluster; and

in accordance with the application-layer availability of a server node in the load-balancing cluster as determined by the application-layer monitoring of the control client, sending to the load-balancing cluster a message that is configured to control whether the server node handles application-layer requests sent to the load-balancing cluster by the user clients, where the application-layer requests sent by the control client to monitor application-layer availability conform to a same application-layer protocol that the user client requests conform to.

21. A method as recited in claim 20, wherein the monitoring comprises selectively deactivating one or more active members of the cluster.

22. A method as recited in claim 20, wherein the monitoring comprises, based upon the monitoring, identifying one or more active members of the cluster that are presently overwhelmed at the application-layer.

23. A method as recited in claim 20, wherein the monitoring comprises:

based upon the monitoring, identifying one or more active members of the cluster that are presently overwhelmed at the application-layer;

exoclusterly deactivating one or more members identified by the identifying.

24. A method as recited in claim 20, wherein the monitoring comprises selectively activating one or more inactive members of the load-balancing cluster.

25. A method as recited in claim 20, wherein the monitoring comprises, based upon the monitoring, identifying one or more inactive members of the cluster that are not presently overwhelmed at the application-layer.

26. A method as recited in claim 20, wherein the monitoring comprises:

based upon the monitoring, identifying one or more inactive members of the cluster that are not presently overwhelmed at the application-layer;

exoclusterly activating one or more members identified by the identifying.

27. A method as recited in claim 20, wherein the monitoring comprises:

based upon the monitoring, identifying one or more active members of the cluster that are presently overwhelmed at the application-layer and identifying one or more inactive members of the cluster that are not presently overwhelmed at the application-layer;

exoclusterly deactivating one or more active members identified by the identifying;

exoclusterly activating one or more inactive members identified by the identifying.

28. A method as recited in claim 20 further comprising determining a present activity state of the members of the cluster.

29. A method as recited in claim 20 further comprising:

determining a present activity state of the members of the cluster;

tracking and persisting the activity states of the members of the cluster.

30. A method as recited in claim 29, wherein the activity state includes a cluster state.

31. A method as recited in claim 29 further comprising reporting a present activity state of one or more members of the cluster.

32. A method as recited in claim 29 further comprising reporting historical record of the activity states of one or more members of the cluster.

33. A method as recited in claim 29 further comprising reporting a present application-layer state of one or more members of the cluster.

34. A method as recited in claim 29 further comprising reporting historical record of the application-layer states of one or more members of the cluster.

35. A method as recited in claim 20, wherein the monitoring comprises monitoring in one or more different application-layer protocols.

36. A computer-readable storage medium having computer-executable instructions that, when executed by a computer, performs the method as recited in claim 20.

37. A computer-readable storage medium having computer-executable instructions that, when executed by a computer, perform a method performed by a control client that communicates with a load-balancing cluster of server nodes configured to service application-layer requests sent by user clients to a virtual address common to the cluster, the method comprising:

dynamically determining, by the control client, which server nodes are members of the load-balancing cluster;

monitoring application-layer availability of the server nodes in the load-balancing cluster, the monitoring being performed by the control client generating and sending, from outside of the load-balancing cluster application-layer, requests to those server nodes determined to be members of the load-balancing cluster; and

in accordance with the application-layer availability of a server node in the load-balancing cluster as determined by the application-layer monitoring of the control client, sending to the load-balancing cluster a message that is configured to control whether the server node handles application-layer requests sent to the load-balancing cluster by the user clients where the application-layer requests sent by the control client to monitor application-layer availability conform to a same application-layer protocol that the user client requests conform to.

38. A system comprising:

a control client that communicates with a load-balancing cluster of server nodes configured to service application-layer requests sent by user clients to a virtual address common to the cluster;

a dynamic cluster-membership determiner configured to exoclusterly and dynamically determine, by the control client, which server nodes are members of the load-balancing cluster;

an exocluster monitor configured to monitor application-layer availability of the server nodes in the load-balancing cluster, the monitoring being performed by the control client generating and sending, from outside of the load-balancing cluster application-layer, requests to those server nodes determined to be members of the load-balancing cluster; and

in accordance with the application-layer availability of a server node in the load-balancing cluster as determined by the application-layer monitoring of the control client sending to the load-balancing cluster a message that is configured to control whether the server node handles application-layer requests sent to the load-balancing cluster by the user clients where the application-layer requests sent by the control client to monitor application-layer availability conform to a same application-layer protocol that the user client requests conform to.

39. A system as recited in claim 38 further comprising an overload-identifier configured to identify, based upon the monitored availability, one or more active members of the cluster that are presently overwhelmed at the application-layer.

40. A system as recited in claim 38 further comprising an overload-identifier configured to identify, based upon the monitored availability, one or more inactive members of the cluster that are not presently overwhelmed at the application-layer.

41. A system as recited in claim 38 further comprising a state-determiner configured to determine a present activity state of members of the cluster.

42. A system as recited in claim 38 further comprising:

a state-determiner configured to determine a present activity state of members of the cluster;

a database configured to store the activity states of the members of the cluster.

43. A system as recited in claim 38, wherein the exocluster monitor is protocol agnostic.

44. A system comprising:

a control client that communicates with a load-balancing cluster of server nodes configured to service application-layer requests sent by user clients to a virtual address common to the cluster;

an exocluster monitor configured to monitor application-layer availability of the server nodes in the load-balancing cluster, the monitoring being performed by the control client generating and sending, from outside of the load-balancing cluster application-layer, requests to those server nodes determined to be members of the load-balancing cluster; and

in accordance with the application-layer availability of a server node in the load-balancing cluster as determined by the application-layer monitoring of the control client, sending to the load-balancing cluster a message that is configured to control whether the server node handles application-layer requests sent to the load-balancing cluster by the user clients, where the application-layer requests sent by the control client to monitor application-layer availability conform to a same application-layer protocol that the user client requests conform to.

45. A system as recited in claim 44, wherein the exocluster monitor is further configured to exoclusterly and selectively deactivate one or more active members of the cluster.

46. A system as recited in claim 44 further comprising an overload-identifier configured to identify, based upon the monitored availability, one or more active members of the cluster that are presently overwhelmed at the application-layer.

47. A system as recited in claim 44, wherein the exocluster monitor is further configured to exoclusterly and selectively activate one or more inactive members of the cluster.

48. A system as recited in claim 44 further comprising an overload-identifier configured to identify, based upon the monitored availability, one or more inactive members of the cluster that are not presently overwhelmed at the application-layer.

49. A system as recited in claim 44 further comprising a state-determiner configured to determine a present activity state of the members of the cluster.

50. A system as recited in claim 44 further comprising:

a state-determiner configured to determine a present activity state of the members of the cluster;

a database configured to store the activity states of the members of the cluster.

51. A system as recited in claim 44, wherein the monitor is protocol agnostic.

52. A dynamic, active, exocluster monitoring system for monitoring application-layer availability of server nodes in a load-balancing cluster and for controlling an activity state of such server nodes, the monitoring system comprising:

a control client that communicates with a load-balancing cluster of server nodes configured to service application-layer requests sent by user clients to a virtual address common to the cluster;

an app-monitor configured to exoclusterly monitor application-layer availability of the server nodes in the load-balancing cluster the, monitoring being performed by the control client generating and sending, from outside of the load-balancing cluster application-layer, requests to those server nodes determined to be members of the load-balancing cluster; and

in accordance with the application-layer availability of a server node in the load-balancing cluster as determined by the application-layer monitoring of the control client, sending to the load-balancing cluster a message that is configured to control whether the server node handles application-layer requests sent to the load-balancing cluster by the user clients, where the application-layer requests sent by the control client to monitor application-layer availability conform to a same application-layer protocol that the user client requests conform to.

53. A system as recited in claim 52 further comprising a database configured to store state change information, including cluster state and application-layer state.

54. A system as recited in claim 52 further comprising multiple app-monitors.

55. A system as recited in claim 52 further comprising multiple cluster-controls.