Dynamic load balancing of a network of client and server computer

• US 20030126200A1
• Filed: 05/13/2002
• Published: 07/03/2003
• Est. Priority Date: 08/02/1996
• Status: Active Grant

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: James J. Wolff
• Inventors: Wolff, James J.
• Granted Patent: US 6,886,035 B2
• Time in Patent Office: Days
• US Class Current: 709/203
• CPC Class Codes: G06F 9/52 Program synchronisation; Mutual exclusion, e.g. by means of semaphores

Abstract

Methods for load rebalancing by clients in a network are disclosed. Client load rebalancing allows the clients to optimize throughput between themselves and the resources accessed by the nodes. A network which implements this embodiment of the invention can dynamically rebalance itself to optimize throughput by migrating client I/O requests from overutilized pathways to underutilized pathways.

Client load rebalancing refers to the ability of a client enabled with processes in accordance with the current invention to remap a path through a plurality of nodes to a resource. The remapping may take place in response to a redirection command emanating from an overloaded node, e.g. server. These embodiments disclosed allow more efficient, robust communication between a plurality of clients and a plurality of resources via a plurality of nodes.

In an embodiment of the invention a method for load balancing on a network is disclosed. The network includes at least one client node coupled to a plurality of server nodes, and at least one resource coupled to at least a first and a second server node of the plurality of server nodes. The method comprises the acts of: receiving at a first server node among the plurality of server nodes a request for the at least one resource; determining a utilization condition of the first server node; and re-directing subsequent requests for the at least one resource to a second server node among the plurality of server nodes in response to the determining act. In another embodiment of the invention the method comprises the acts of: sending an I/O request from the at least one client to the first server node for the at least one resource; determining an I/O failure of the first server node; and re-directing subsequent requests from the at least one client for the at least one resource to an other among the plurality of server nodes in response to the determining act.

Claims

1. A method for load balancing on a network, the network including at least one client node coupled to a plurality of server nodes, and at least one resource coupled to at least a first and a second server node of the plurality of server nodes, and the method comprising the acts of:

receiving at a first server node among the plurality of server nodes a request for the at least one resource;

determining a utilization condition of the first server node; and

re-directing subsequent requests for the at least one resource to a second server node among the plurality of server nodes in response to the determining act.

2. The method of claim 1, wherein the request is for a data set on the at least one resource.

3. The method of claim 1, wherein an origin of the request is one of, the at least one client node and the second server node.

4. The method of claim 1, wherein the determining act is performed on at least one of:

the at least one client node, the first server node and the second server node.

5. The method of claim 1, wherein the re-directing act is performed on at least one of:

the at least one client node, the first server node and the second server node.

6. The method of claim 1, wherein the determining act further comprises the acts of:

indicating the utilization condition when a utilization of the first server node exceeds a limit; and

sending to the origin of the request a redirect command.

7. The method of claim 6, wherein the limit comprises at least one of a fixed threshold utilization and a calculated threshold utilization.

8. The method of claim 1, wherein the re-directing act further comprises the acts of:

receiving at the origin of the request the redirect command;

discovering that the second server node couples to the at least one resource;

ascertaining the utilization condition by detecting that a utilization of the first server node exceeds the utilization of the second server node;

re-directing subsequent requests for the at least one resource from the origin to the second server node, responsive to the ascertaining act; and

sending to the origin of the request a redirect command.

9. The method of claim 1, wherein the determining act further comprises the act of:

discovering that the second server node couples to the at least one resource;

ascertaining the utilization condition by detecting that the utilization of the first server node exceeds the utilization of the second server node; and

wherein the re-directing act further comprises the acts of;

sending to the origin a redirect command and a path indicating the second server node as the path to the at least one resource; and

re-directing subsequent requests for the at least one resource from the origin to the second server node, responsive to the sending act.

10. The method of claim 1, wherein the determining act further comprises the acts of:

detecting at the origin a delay in a response from the first server node to the request for the at least one resource.

11. The method of claim 10, wherein the re-directing act further comprises the act of:

detecting that the second server node couples to the at least one resource;

polling the second among the plurality of server nodes to ascertain that the utilization of the second server node is less than a limit; and

wherein the re-directing act further comprises the act of,

re-directing subsequent requests from the origin for the at least one resource to the second server node.

12. The method of claim 11, wherein the limit comprises at least one of a utilization level less than the first server node and a utilization less than a fixed amount.

13. The method of claim 1, wherein the re-directing act further comprises the acts of:

tracking each of the plurality of server nodes to determine the utilization for each of the plurality of server nodes;

detecting that the second server node couples to the at least one resource;

ascertaining that the utilization of the second among the plurality of server nodes is less than the utilization of the first of the plurality of server nodes; and

re-directing subsequent requests from the origin for the at least one resource to the second server node.

14. The method of claim 13, wherein the tracking act further comprises the acts of:

determining at each of the plurality of server nodes a corresponding utilization level;

updating from each of the plurality of server nodes a corresponding entry in a file for storing utilization levels for the network; and

wherein the ascertaining act further comprises the act of;

reading the file to detect that the utilization of the second server node is less than the utilization of the first server node.

15. The method of claim 14, wherein the updating act further comprises:

periodically updating the corresponding entry in the file at an interval recorded in the file.

16. The method of load balancing of claim 1, wherein a measure from which the utilization condition is determined includes processor utilization and I/O utilization.

17. The method of claim 1, further comprising the acts of:

forming a single system image at the at least one client node showing all available paths through corresponding available ones of the plurality of server nodes, including the first server node, to the at least one resource.

18. The method of claim 17, wherein the re-directing act further comprises the acts of:

correlating in the single system image, the at least one resource with all corresponding ones of the plurality of server nodes coupled to the at least one resource;

tagging each of the corresponding available ones with a time tag indicating the time at which a corresponding redirect command was last received by the at least one client node from the corresponding one; and

selecting the second server node as a path to the at least one resource based on a comparison of the time tags indicating that the second server node is the least recently rebalanced.

19. The method of claim 18, wherein the updating act further comprises the act of:

updating the single system image to exclude from among corresponding available ones of the plurality of server nodes, a one of the plurality of server nodes which becomes unavailable; and

to include among the corresponding available ones of the plurality of server nodes the one when it becomes available.

20. A method for load balancing on a network, the network including at least one client node coupled to a plurality of server nodes, and at least one resource coupled to at least a first and a second server node of the plurality of server nodes, and the method comprising the acts of:

sending an I/O request from the at least one client to the first server node for the at least one resource;

determining an I/O failure of the first server node;

re-directing subsequent requests from the at least one client for the at least one resource to an other among the plurality of server nodes in response to the determining act.

21. The method of claim 20 further comprising the acts of:

forming a single system image at the at least one client node showing all available paths through corresponding available ones of the plurality of server nodes, including the first server node, to the at least one resource; and

updating the single system image to exclude from among corresponding available ones of the plurality of server nodes, the first server node responsive to the determining act.

22. The method of 21 further comprising the act of:

updating the single system image to include among the corresponding available ones of the plurality of server nodes the first server node when it becomes available.

23. The method of claim 22 further comprising the acts of:

correlating in the single system image, the at least one resource with all corresponding ones of the plurality of server nodes coupled to the at least one resource;

tagging each of the corresponding available ones with a time tag indicating the time at which a corresponding redirect command was last received by the at least one client node from the corresponding one; and

selecting the second server node as a path to the at least one resource based on a comparison of the time tags indicating that the second server node is the least recently rebalanced.

24. A method for load balancing on a network, the network including, at least one client node coupled to a plurality of server nodes and at least a first and a second resource coupled to respectively a first and a second server node among the plurality of server nodes;

and the method comprising the acts of;

receiving at the first server node a request from the at least one client node for the first resource;

determining a utilization condition on the first of the plurality of server nodes; and

redirecting subsequent requests for the first resource to the second resource via the second server node based on a determination that the first and second resources offer similar features and in response to the determining act.

25. A computer usable medium having computer readable program code means embodied therein for load balancing on a network, the network including, at least one client node coupled to a plurality of server nodes and at least one resource coupled to at least a first and a second server node of the plurality of server nodes, and the computer readable program code means in said article of manufacture comprising:

computer readable program code means for receiving at a first of the plurality of server nodes a request for the at least one resource;

computer readable program code means for determining a utilization condition on the first server node; and

computer readable program code means for re-directing subsequent requests for the at least one resource from the at least one client node to an other among the plurality of server nodes in response to the determining act.