Load balancing

• US 7,984,148 B2
• Filed: 11/10/2003
• Issued: 07/19/2011
• Est. Priority Date: 07/15/1998
• Status: Expired due to Term

• Current Assignee: Radware Limited
• Original Assignee: Radware Limited
• Inventors: Zisapel, Roy, Peless, Amir
• Primary Examiner(s): Coulter; Kenneth R
• Application Number:
  US10/703,540
• Publication Number:
  US 20050022203A1
• Time in Patent Office: 2,808 Days
• Field of Search: 718/105, 718/104, 718/102, 709/226, 709/238, 709/224
• US Class Current: 709/226
• CPC Class Codes: G06F 9/505 considering the load
• CPC Class Codes: H04L 61/4511 using domain name system [DNS]
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1006 with static server selection, e.g. the same server being selected for a specific client
• 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 69/32 Architecture of open systems interconnection [OSI] 7-layer type protocol stacks, e.g. the interfaces between the data link level and the physical level

Abstract

A method for load balancing requests on a network, the method including receiving a request from a requester having a requester network address at a first load balancer having a first load balancer network address, the request having a source address indicating the requestor network address and a destination address indicating the first load balancer network address, forwarding the request from the first load balancer to a second load balancer at a triangulation network address, the request source address indicating the requester network address and the destination address indicating the triangulation network address, the triangulation network address being associated with the first load balancer network address, and sending a response from the second load balancer to the requestor at the requestor network address, the response having a source address indicating the first load balancer network address associated with the triangulation network address and a destination address indicating the first requestor network address.

Claims

1. A method of load balancing in a network comprising multiple geographic network locations, each of at least two of such locations respectively having a load balancer, the load balancer at one of said locations being referred to as a first load balancer and the load balancer at the other of said locations being referred to as a second load balancer, each of said load balancers having a respective IP address and each load balancer serving as a gateway to respective servers at that location, the method comprising the steps of:

polling by at least said first and second load balancers of a polled client located at yet an other network location different from the respective locations of said at least first and second load balancers to determine a hop count related measurement or a round trip latency related measurement, respectively, between each of said at least first and second load balancers and said polled client;

collecting by said first load balancer of the measurement information from both said at least first and second load balancers;

maintaining by said first load balancer traffic load or capacity information related to servers at its location as well as related to said second load balancer;

receiving by said first load balancer of an HTTP request having a destination IP address from a requesting client at said other network location;

performing by said first load balancer a selection between said servers at its own location and the at least said second load balancer for servicing said HTTP request, and

forwarding of said HTTP request by said first load balancer to at least said second load balancer using the IP address of said at least second load balancer, as the destination IP address of the forwarded HTTP request, when said first load balancer selects said second load balancer to service said HTTP request.

2. A method according to claim 1, wherein said requesting client is a client other than the polled client that was, both clients however being in a same IP subnet.

3. A method according to claim 1, wherein said polled client has an IP address and polling of a client is achieved by using a polling request having an destination IP address and using the IP address of the polled client as the destination IP address in the polling request from said at least first and second load balancers.

4. A method according to claim 1, wherein said selection is determined by said first load balancer based upon said hop count measurements or round trip latency measurements information as well as the information about the traffic load or processing capacity.

5. A method according to claim 1, wherein said first load balancer stores the information on traffic load or capacity information of its servers and of said at least second load balancer, as well as the measurement information polled by both said at least first and second load balancers, in a storage memory in said first load balancer.

6. A network load balancing system, comprising:

a network;

at least first and second load balancing devices, said at least first and second load balancing devices each having an IP address and each being located at geographically different network locations;

a plurality of servers, each load balancing device serving as a gateway to respective ones of said servers that it services;

a storage memory in said first load balancing device for storing traffic load or processing capacity information related to its servers and to at least said second load balancing device;

a plurality of clients;

each of said at least first and second load balancing devices are operative to poll a first client located at an other network location different from the locations of said at least first and second load balancing devices, to determine the hop count related measurement or the round trip latency related measurement, respectively, between each of said at least first and second load balancing devices and said first client;

said first load balancing device is operative to collect the measurement information from both said at least first and second load balancing devices and storing the information in said storage memory;

said first load balancing device is operative to receive an HTTP request having a destination IP address from a second client, and based upon the information stored in said storage memory make a selection between its servers and said at least second load balancing device for servicing said HTTP request; and

when selecting said second load balancing device, said first load balancing device is operative to forward said HTTP request to the second load balancing device using the IP address of the second load balancing device as the destination IP address of the forwarded HTTP request.

7. The network according to claim 6, wherein said first client and second client are the same one.

8. The network according to claim 6, wherein said first and second clients each have a respective IP address and wherein said IP addresses are different, however both being in the same IP subnet.

9. A network according to claim 6, wherein said first and second clients each have a respective IP address, wherein said at least first and second load balancing devices are operative to poll said first client by using a polling request having a destination IP address and using the IP address of the first client as the destination IP address of the polling request from said at least first and second load balancing devices.

10. A load balancing device for load balancing at least two servers located within a computer network, the network comprising multiple geographic network locations with at least another load balancing device at another network location, said load balancing device comprising:

a storage memory located therein for storing traffic load or processing capacity information related to its servers and to at least the another load balancing device;

a polling mechanism for polling a polled client located at a network location different from the network location of said load balancing device and different from the network location of the another load balance device, the polling mechanism determines the hop count related measurement or the round trip latency related measurement, between said load balancing device and said polled client;

a collector for requesting the another load balancing device to obtain similar information between that another load balancing device and the same client and for collecting the measurement information obtained from said polling mechanism as well as the information the another load balancing device obtains, and storing all such information in said storage memory;

a receiving mechanism for receiving an HTTP request having a destination IP address from a requesting client;

a selector operative to make a selection between using its servers or said at least another load balancing device for servicing said HTTP request based upon the information stored in said storage memory; and

a transmitter for forwarding the HTTP request to the at least the another load balancing device using the IP address of the said another load balancing device as the destination IP address of the forwarded HTTP request, when it selects the another load balancing device to service the HTTP request.

11. The device according to claim 10, wherein said HTTP request is from the same client that was polled.

12. The network according to claim 10, wherein said HTTP request is from a requesting client other then the polled client, both clients however being in a same IP subnet.

13. A device according to claim 10, wherein said polled client has an IP address, and said polling mechanism polls said polled client by using a polling request having a destination IP address and using the IP address of the polled client as the destination IP address of the polling request.