QoS based load-balance policy for WLAN

• US 20050053046A1
• Filed: 09/10/2004
• Published: 03/10/2005
• Est. Priority Date: 09/10/2003
• Status: Active Grant

• Current Assignee: Delta Networks Incorporated (Delta Electronics Incorporated)
• Original Assignee: Delta Networks Incorporated (Delta Electronics Incorporated)
• Inventors: Wang, Shiwei
• Granted Patent: US 7,675,890 B2
• Time in Patent Office: Days
• US Class Current: 370/338
• CPC Class Codes: H04L 12/4645 Details on frame tagging routing of packets H04L45/00; support for virtual LAN H04L49/354
• CPC Class Codes: H04L 47/10 Flow control; Congestion control
• CPC Class Codes: H04L 47/125 by balancing the load, e.g. traffic engineering
• CPC Class Codes: H04L 47/2408 for supporting different services, e.g. a differentiated services [DiffServ] type of service
• CPC Class Codes: H04L 47/2433 Allocation of priorities to traffic types
• CPC Class Codes: H04L 47/2491 Mapping quality of service [QoS] requirements between different networks
• CPC Class Codes: H04W 28/0827 Triggering entity
• CPC Class Codes: H04W 28/0862 of same hierarchy level
• CPC Class Codes: H04W 28/0942 based on measured or predicted load of entities- or links
• CPC Class Codes: H04W 28/0983 for optimizing bandwidth or throughput
• CPC Class Codes: H04W 28/10 Flow control between communication endpoints
• CPC Class Codes: H04W 40/02 Communication route or path selection, e.g. power-based or shortest path routing
• CPC Class Codes: H04W 8/04 Registration at HLR or HSS [Home Subscriber Server]
• CPC Class Codes: H04W 84/12 WLAN [Wireless Local Area Networks]
• CPC Class Codes: Y02D 30/70 in wireless communication networks

Abstract

A load balancing method for a wireless local area network (LAN) which has a plurality of access points is provided in the present invention. The load balancing decision is made by a load balancing module according to traffic conditions and bandwidth availability of each traffic priority class based on a corresponding class of service. The load balancing module is a centralized module which the load balancing decision is determined by the centralized module, or is determined through information exchanges among distributed load balancing modules of the access points. The load balancing decision also takes into consideration of VLANs per VLAN tag basis. Each of the VLANs is defined with the corresponding traffic priority class.

Claims

1. A load balancing method for a wireless local area network (LAN) having a plurality of access points, wherein the load balancing decision is made by a load balancing module according to traffic conditions and bandwidth availability of each traffic priority class based on a corresponding class of service.

2. The load balancing method as claimed in claim 1, wherein the load balancing module is a centralized module which the load balancing decision is determined by the centralized module.

3. The load balancing method as claimed in claim 1, wherein the load balancing decision is determined through information exchanges among distributed load balancing modules of the access points.

4. The load balancing method as claimed in claim 1, wherein the load balancing decision being made also by taking into consideration of a plurality of virtual local area networks (VLANs) per VLAN tag basis.

5. The load balancing method as claimed in claim 4, wherein each of the VLANs is defined with the corresponding traffic priority class.

6. The load balancing method as claimed in claim 5, wherein when the corresponding traffic priority class of the VLAN tag is mapped to a constant-bit-rate traffic type, the load balancing decision being made by the load balancing module of one of the access point comprising:

reserving and dividing a bandwidth for each of the access points in according to a bit rate to get a number of concurrent connections that are accommodated under reservation; and

a new constant-bit-rate request being admitted if and only if a quantum of bandwidth equals to the bit-rate is available, and the quantum of bandwidth is assigned to a requesting session.

7. The load balancing method as claimed in claim 4, wherein each of the access points having a plurality of queues.

8. The load balancing method as claimed in claim 7, wherein each of the queues is assigned the plurality of corresponding classes of service, each of the class of service being mapping to the plurality of virtual local area networks (VLANs), each of a plurality of service set identifiers (SSIDs) being mapping to a VLAN tag corresponding to one of the VLANs.

9. The load balancing method as claimed in claim 8, wherein the relationship between SSID and VLAN tag is one-to-one.

10. The load balancing method as claimed in claim 8, wherein the relationship between SSID and VLAN tag is many-to-one.

11. The load balancing method as claimed in claim 8, wherein the load balancing module performs load balancing in according to the corresponding traffic priority class within the same VLAN and the same SSID.

12. The load balancing method as claimed in claim 8, wherein the load balancing module performs load balancing in according to the corresponding traffic priority class within the same VLAN and the plurality of SSIDs.

13. The QoS load balancing method as claimed in claim 7, wherein each of the queues is assigned the plurality of corresponding classes of service, each of the class of service being mapping to the plurality of virtual local area networks (VLANs), each of a plurality of basic service set identifier (BSSID) being mapping to a VLAN tag corresponding to one of the VLANs.

14. The load balancing method as claimed in claim 13, wherein the relationship between BSSID and VLAN tag is one-to-one.

15. The load balancing method as claimed in claim 13, wherein the relationship between BSSID and VLAN tag is many-to-one.

16. The load balancing method as claimed in claim 13, wherein the load balancing module performs load balancing in according to the corresponding traffic priority class within the same VLAN and the same BSSID.

17. The load balancing method as claimed in claim 13, wherein the load balancing module performs load balancing in according to the corresponding traffic priority class within the same VLAN and the BSSIDs.

18. The load balancing method as claimed in claim 7, wherein when the corresponding traffic priority class of the VLAN tag is mapped to a variable-bit-rate traffic type, the load balancing decision being made by the load balancing module of one of the access point comprising using a tuple of a congestion level indicator (CLI) to indicate if congestion occurs.

19. The load balancing method as claimed in claim 18, wherein the load balancing decision is made by the load balancing module further comprises revoking reverse roaming to force de-association and sequentially perform re-association of traffics to achieve more optimized traffic distribution.

20. The load balancing method as claimed in claim 18, wherein the tuple has a plurality of attributes to indicate if the congestion occurs.

21. The load balancing method as claimed in claim 20, wherein the attributes of the tuple are a queue length, utilization level or the number of concurrent sessions.

22. The load balancing method as claimed in claim 20, wherein the attributes of the tuple comprises reception of various clients, the reception of the clients comprises a link quality for each of the clients or a signal noise ratio (S/N ratio).

23. The load balancing method as claimed in claim 20, wherein each of the attributes is associated with a set of high and low water marks.

24. The load balancing method as claimed in claim 20, wherein the load balancing decision is made in according to one of the attributes of the tuple independently.

25. The load balancing method as claimed in claim 20, wherein the load balancing decision is made by a policy-based trigger, which is defined based on a first logic expression of a portion or all of the attributes of the tuple.

26. The load balancing method as claimed in claim 25, wherein the first logic expression between the attributes of the tuple is any combination or equivalents of AND/OR relationships of the attributes of the tuple, or equivalence which is expressed by the AND/OR relationships of the attributes.

27. The load balancing method as claimed in claim 25, wherein each of the attributes is selectively associated with a set of high and/or low water marks and the policy-based trigger is defined selectively based on any combination of a second logic expression of the high and/or low water marks of the attributes of the tuple.

28. The load balancing method as claimed in claim 27, wherein the second logic expression of the high and/or low water marks of the attributes of the tuple is any combination or equivalents of AND/OR relationships of the high and/or low water marks of the attributes, or equivalence which is expressed by the AND/OR relationships of the attributes.

29. The load balancing method as claimed in claim 25, wherein the policy-based trigger is activated when one or a portion of the attributes of the tuple reaches the high water mark.

30. The load balancing method as claimed in claim 25, wherein the policy-based trigger is activated when a portion or all of the attributes of the tuple drop to the low water mark.

31. The load balancing method as claimed in claim 25, wherein each of the high water marks and the low water marks of the attributes of the tuple is assigned its corresponding weight, the policy-based trigger is defined further based one the first logic expression of a portion or all of the weights of the high water marks and the low water marks.

32. The load balancing method as claimed in claim 31, wherein the weights of the high water marks and the low water marks of the attributes of the tuple is determined based on a cost function.

33. The load balancing method as claimed in claim 32, wherein the weights of the high water marks and the low water marks are selectively adaptive.

34. The load balancing method as claimed in claim 33, wherein the weights of the high water marks and the low water marks are selectively adaptive in according to traffic conditions between the APs.

35. The load balancing method as claimed in claim 34, wherein the weights of the high water marks and the low water marks are adaptive in according to conditions defined by network administrators.

36. The load balancing method as claimed in claim 35, wherein the results of the cost functions can be selected in according to an optimal condition set up by the network administrators.

37. The load balancing method as claimed in claim 32, wherein the result of the cost function is a binary decision, which depends on the relationship of the value of the attribute and the high water mark or the low water mark.

38. The load balancing method as claimed in claim 37, wherein the value of the attribute is higher than the high water mark, the weight corresponding to the attribute is used for the cost function.

39. The load balancing method as claimed in claim 32, wherein the result of the cost function is a quantification decision, which depends on the differences between the value of the attribute and the high water mark in one or more of the attributes of the tuple.

40. The load balancing method as claimed in claim 39, wherein the quantification decision depends on a plurality of normalization values which is achieved by a percentage value obtained by comparing the difference with the value of high water mark.

41. The load balancing method as claimed in claim 39, wherein the quantification decision depends on a plurality of normalization values which is achieved by a percentage value obtained by comparing the difference with a value set by the network administrators.

42. The load balancing method as claimed in claim 32, wherein the result of the cost function is a quantification decision, which depends on the differences between the value of the attribute and the low water mark in one or more of the attributes of the tuple.

43. The load balancing method as claimed in claim 42, wherein the quantification decision depends on a plurality of normalization values which is achieved by a percentage value obtained by comparing the difference with the value of high water mark.

44. The load balancing method as claimed in claim 43, wherein the quantification decision depends on a plurality of normalization values which is achieved by a percentage value obtained by comparing the difference with a value set by the network administrators.

45. The load balancing method as claimed in claim 25, wherein the policy-based trigger is selected from a set of default policies which are predefined in the load balancing module and selected by a network administrator.

46. The load balancing method as claimed in claim 25, wherein the trigger is activated for onset of congestion, the load balance module examines the tuples of the CLIs in all of the overlapping access points with priority levels of the traffic priority classes higher than or equal to the overloaded traffic stream and selects traffic sessions corresponding to the SSIDs or BSSIDs with CLIs based on conditions set by the trigger as targets for shifting the load over.

47. The load balancing method as claimed in claim 46, wherein the traffic sessions corresponding to SSIDs or BSSIDs with CLIs below the high water mark are selected according to the trigger as target for shifting the load over.

48. The load balancing method as claimed in claim 46, wherein the the SSIDs or BSSIDs of the selected traffic sessions are all mapped to the same VLAN.

49. The load balancing method as claimed in claim 46, wherein when the traffic sessions corresponding to the SSIDs or BSSIDs with CLIs based on the conditions set by the trigger as targets for shifting the load over, the load is moved across these selected SSIDs or BSSIDs in all of the overlapping access points and the loads of the SSIDs or BSSIDs within the same VLAN are roughly the same.

50. The load balancing method as claimed in claim 25, wherein when the CLI of one of the SSIDs or BSSIDs in the same VLAN drops below the low watermark, the load balance module moves over the traffics from other SSIDs or BSSIDs in the same VLAN in all of the overlapping access points until the CLIs of the SSIDs or BSSIDs in the same VLAN become roughly the same.

51. The load balancing method as claimed in claim 50, wherein the step of moving over the traffic sessions by the load balance module comprises moving the traffic sessions of the SSIDs or BSSIDs above the low water mark to the SSIDs or BSSIDs with CLIs lower than the low water mark until the CLIs of the SSIDs or BSSIDs in the same VLAN become roughly the same.

52. The load balancing method as claimed in claim 25, wherein to balance the traffic loads, the sessions in sleep modes are selected first for load balancing.

53. The load balancing method as claimed in claim 25, wherein to balance the traffic loads, reverse roaming is applied to activate the variable-bit-rate traffics.

54. The load balancing method as claimed in claim 25, wherein the policy used for the load balancing is independent from the choice of a service policy.

55. The load balancing method as claimed in claim 54, wherein the service policy is First-Come, First-Served (FCFS), a strict priority, or a weighted fair-queuing.

56. A scheme for load balancing in a wireless local area network (LAN), the LAN having a plurality of access points, each of the access point in the scheme comprising a Hostapd program, the Hostapd program having a tuple of a congestion level indicator, the tuple having a plurality of attributes, the Hostapd program executing load balancing by triggering a policy based on the attributes in the tuple of the congestion level indicator.

57. The scheme for load balancing in a wireless local area network (LAN) of claim 56, wherein the Hostapd program comprises a load balance module for receiving a queue statistic information, wherein the congestion level indicator is updated with a queue length information from the queue statistic information.

58. The scheme for load balancing in a wireless local area network (LAN) of claim 56, wherein the Hostapd program comprises a bandwidth monitor module, for sampling statistical information of a bandwidth, wherein the congestion level indicator is updated with utilization information from the bandwidth monitor module.

59. The scheme for load balancing in a wireless local area network (LAN) of claim 56, wherein the congestion level indicator is transmitted to a wireless driver and sets the wireless driver for executing actions based on the updated congestion level indicator.

60. The scheme for load balancing in a wireless local area network (LAN) of claim 56, wherein the Hostapd program is configured by a configuration system program.

61. The scheme for load balancing in a wireless local area network (LAN) of claim 56, wherein the Hostapd program is configured by a Hostapd.conf file stored in the configuration system program.

62. The scheme for load balancing in a wireless local area network (LAN) of claim 61, wherein the Hostapd.conf file is created by the configuration system program.