Method for integrated load balancing among peer servers

• US 20030158940A1
• Filed: 02/20/2002
• Published: 08/21/2003
• Est. Priority Date: 02/20/2002
• Status: Active Grant

• Current Assignee: Valtrus Innovations Limited (f/k/a Dolya Holdco 9 Limited) (Key Patent Innovations Limited)
• Original Assignee: Hewlett-Packard Development Company, L.P. (HP Inc.)
• Inventors: Leigh, Kevin B.
• Granted Patent: US 7,284,067 B2
• Time in Patent Office: Days
• US Class Current: 709/226
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/101 based on network conditions
• CPC Class Codes: H04L 67/1012 based on compliance of requirements or conditions with available server resources
• CPC Class Codes: H04L 67/1029 using data related to the state of servers by a load balancer

Abstract

A system for balancing network loads among a group of computer servers. A load balancing module is provided for each server and includes multiple network ports for receiving network traffic and coupling messages to its server or to another module. The interconnect map for the modules is automatically determined by a topology discovery system and maintained by a topology update system. Each module monitors the loading and health of its server and provides this information to the other modules. One or more master modules are directly connected to an external network and route received traffic to their own servers and other servers through other modules or a networking switch based on a priority system. The priority system may be based on dynamic conditions of server capacities and current loading.

Claims

1. Apparatus for balancing network traffic among a group of servers, comprising:

a plurality of load balancing modules, one coupled to each server and each having at least two network ports, and

an internal network of network segments connecting at least one network port of each load balancing module to a network port of at least one other load balancing module in said plurality of load balancing modules, and coupling every load balancing module, directly or through one or more load balancing modules, to each of the other load balancing modules,

at least one of said load balancing modules having a network port coupled to an external network.

2. Apparatus according to claim 1, wherein:

two load balancing modules have a network port directly coupled to said external network.

3. Apparatus according to claim 1, wherein:

said plurality of load balancing modules comprises two load balancing modules,

each of said two load balancing modules has two network ports,

one of said two network ports of each load balancing module is coupled directly to said external network, and

the other of said two network ports of said load balancing modules are coupled together.

4. Apparatus according to claim 1, wherein:

said plurality of load balancing modules comprises at least three load balancing modules,

each of said load balancing modules has three network ports, and

two of said load balancing modules have one network port coupled directly to said external network.

5. Apparatus according to claim 1, wherein:

at least one load balancing module which is directly coupled to the external network receives external network messages and routes messages to its own server or through said internal network to servers coupled to said other load balancing modules according to a priority system.

6. Apparatus according to claim 5, wherein:

said priority system is based on workload levels of said servers.

7. Apparatus according to claim 5, wherein:

said priority system is based on workload capacities of said servers.

8. Apparatus according to claim 5, wherein:

each module monitors workload conditions of the server to which it is coupled and communicates said conditions to other modules.

9. Apparatus according to claim 5, wherein:

two load balancing modules are directly coupled to the external network and each is adapted to receive network messages and route messages to its own server or through said internal network to servers coupled to said other load balancing modules according to a priority system.

10. Apparatus according to claim 9, wherein:

said two directly external network coupled modules receive redundant external network messages,

a first of said two directly external network coupled modules routes messages to its own server or through other load balancing modules to servers coupled to said other load balancing modules according to a priority system, and

if said first of said two directly external network coupled modules fails to operate correctly, a second of said two directly external network coupled modules routes messages to its own server or through said internal network to servers coupled to said other load balancing modules according to a priority system.

11. Apparatus according to claim 1, further comprising:

means in each load balancing module for sending messages to and receiving messages from other load balancing modules through said internal network.

12. Apparatus according to claim 11, further comprising:

means in each module for identifying the topology of the internal network.

13. Apparatus according to claim 12, further comprising:

means in each load balancing module for detecting a change in the topology of the internal network and sending a message identifying said change over said internal network to other load balancing modules.

14. Apparatus according to claim 13, further comprising:

means in each load balancing module for storing a map of said topology, and

means for routing a message to a selected load balancing module by the shortest path through said internal network.

15. Apparatus according to claim 14, further comprising:

means in each load balancing module for updating said map of said topology in response to changes detected in said topology.

16. Apparatus according to claim 1, further comprising:

a switch having a plurality of ports,

at least one of said switch ports coupled to at least one port of each of said plurality of load balancing modules,

whereby said plurality of load balancing modules are coupled together through the switch to form said internal network.

17. Apparatus according to claim 16, wherein:

at least one switch port is coupled to said external network.

18. A method for balancing network loads directed to a group of servers, comprising:

using a group of load balancing modules, each coupled to one server in said group of servers and through an internal network to each of the other load balancing modules, to couple external network messages to each server, and

using a first of said load balancing modules to receive an external network message and route the message through said internal network to a server selected according to a priority system.

19. The method of claim 18 further comprising:

routing the message to the server to which said first of said load balancing modules is coupled.

20. The method of claim 18, further comprising:

routing the message through said internal network and one or more of said other load balancing modules to a server other than the server to which said first of said load balancing modules is coupled.

21. The method of claim 18, further comprising:

coupling workload capacity and level information from each of said servers through said internal network and said load balancing modules to said first of said load balancing modules.

22. The method of claim 21, further comprising:

using workload capacity information to selectively route network messages to said servers.

23. The method of claim 21, further comprising:

using workload level information to selectively route external network messages to said servers.

24. The method of claim 23, further comprising:

upon detection of a low workload level, selectively deactivating one or more servers and routing all external network messages to other servers.

25. The method of claim 24, further comprising:

upon detection of an increased workload level, selectively reactivating a previously deactivated server.

26. The method of claim 18, further comprising:

assigning an order to said group of load balancing modules and sequentially routing said messages to said modules in accordance with said order.

27. The method of claim 18, further comprising:

automatically detecting the topology of said internal network.

28. The method of claim 27, further comprising:

transmitting a roll call query message from each port of each load balancing module.

29. The method of claim 28, further comprising:

in response to receipt of a roll call query message at a first load balancing module port, transmitting a roll call response identifying the first load balancing module, the port from which the response is sent, and other ports of said first load balancing module.

30. The method of claim 29, further comprising:

transmitting a roll call response identifying other load balancing modules to which ports of said first load balancing modules are coupled.

31. The method of claim 30, further comprising:

storing a map of said topology in each load balancing module, and

using said topology to route network loads.

32. The method of claim 31, further comprising:

detecting changes in said topology and updating said map.

33. The method of claim 32, further comprising:

sending a query over said internal network from said first of said load balancing modules to each of the other of said load balancing modules to verify network topology.

34. The method of claim 32, further comprising:

detecting in each load balancing module a change in topology and sending over the internal network a message to said first of said load balancing modules indicating the change.

35. The method of claim 34, further comprising:

changing the topology map stored in said first of said load balancing modules to reflect the change, and

transmitting a message over said internal network from said first of said load balancing modules to each of the other of said load balancing modules to update the topology map stored in each of said other of said load balancing modules.

36. A system for load balancing network loads among a group of servers, comprising:

a group of load balancing module means, one coupled to each server in the group of servers, for receiving an external network message and coupling the message to the server to which it is coupled or to another load balancing module means in accordance with a priority system, and

priority means for selecting a server which is to receive the next external network message.

37. The system of claim 36, further comprising:

load reporting means in each load balancing module for monitoring network loading of the server to which the module is coupled and coupling the loading information to other load balancing modules.

38. The system of claim 37 wherein:

the priority system is based on the level of loading of servers.

39. The system of claim 38, wherein:

the priority system is independent of the network protocol layer type.