Global server load balancing

• US 7,581,009 B1
• Filed: 04/27/2007
• Issued: 08/25/2009
• Est. Priority Date: 09/26/2000
• Status: Expired due to Fees

• Current Assignee: Avago Technologies International Sales Pte Limited (Broadcom, Inc.)
• Original Assignee: Foundry Networks Incorporated (Broadcom, Inc.)
• Inventors: Hsu, Ivy Pei-Shan, Cheung, David Chun-Ying, Jalan, Rajkumar Ramniranjan
• Primary Examiner(s): SALAD, ABDULLAHI ELMI
• Application Number:
  US11/741,480
• Time in Patent Office: 851 Days
• Field of Search: 709223-224, 709/226, 709/241, 709/201, 709/203
• 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 67/1012 based on compliance of requirements or conditions with available server resources
• CPC Class Codes: H04L 67/1021 based on client or server locations

Abstract

A global server load balancing (GSLB) switch serves as a proxy to an authoritative DNS communicates with numerous site switches which are coupled to host servers serving specific applications. The GSLB switch receives from site switches operational information regarding host servers within the site switches neighborhood. When a client program requests a resolution of a host name, the GSLB switch, acting as a proxy of an authoritative DNS, returns one or more ordered IP addresses for the host name. The IP addresses are ordered using metrics that include the information collected from the site switches. In one instance, the GSLB switch places the address that is deemed “best” at the top of the list.

Claims

1. A method of performing load balancing using a load balancing switch, the method comprising:

storing, at the load balancing switch, performance metrics regarding a plurality of site switches, each site switch associated with one or more host servers, wherein the one or more host servers associated with a site switch are reachable via the site switch using a virtual address configured at the site switch and wherein the performance metrics regarding a site switch comprise one or more metrics related to one or more host servers associated with the site switch;

storing, at the load balancing switch, a plurality of network addresses generated in response to a domain name query;

ordering the plurality of network addresses at the load balancing switch based upon a first set of performance metrics from the stored performance metrics to generate a first ordered list of network addresses;

determining if a single network address tops the first ordered list generated based upon the first set of performance metrics; and

upon determining that a single network address does not top the first ordered list, reordering, at the load balancing switch, one or more network addresses from the plurality of network addresses based upon a second set of performance metrics from the stored performance metrics, wherein the second set of performance metrics is different from the first set of performance metrics.

2. The method of claim 1 wherein the ordering comprises:

determining at least one network address from the plurality of network addresses that fails a health check; and

excluding the at least one network address from the first ordered list of network addresses.

3. The method of claim 2 wherein the health check is a Open System Interconnection (OSI) layer 4 or layer 7 health check.

4. The method of claim 2 wherein the health check is a Transmission Control Protocol (TCP) or User Datagram Protocol (UDP) health check.

5. The method of claim 1 wherein the ordering or reordering comprises ordering the plurality of network addresses based upon, for each site switch in the plurality of site switches, a number of sessions supported by the site switch and a maximum number of sessions supportable by the site switch.

6. The method of claim 1 wherein storing the performance metrics comprises:

storing the performance metrics in a database stored by the load balancing switch.

7. The method of claim 1 wherein:

the stored performance metrics comprise round trip time information for each site switch in the plurality of site switches and a client machine sending the domain name query, wherein the round trip time information for a site switch is indicative of time for exchanging messages between the site switch and the client machine; and

the ordering or reordering comprises ordering the plurality of network addresses at the load balancing switch based upon the round trip time information included in the performance metrics.

8. The method of claim 1 wherein the ordering or reordering comprises ordering the plurality of network addresses based upon available session capacity associated with network addresses in the plurality of network addresses.

9. The method of claim 1 wherein the ordering or reordering comprises ordering the plurality of network addresses at the load balancing switch based upon a flashback speed associated with each site switch in the plurality of site switches, wherein a flashback speed for a site switch is indicative of a time required for the site switch to respond to a health check performed by the load balancing switch on the site switch.

10. The method of claim 1 wherein the ordering or reordering comprises ordering the plurality of network addresses based upon geographical locations of the host servers associated with the plurality of site switches.

11. The method of claim 1 wherein the ordering or reordering comprises placing a network address from the plurality of network addresses of a least recently selected host server for placement at a higher position in the first or second ordered list.

12. A method performed by a load balancing switch, the method comprising:

storing, at the load balancing switch, performance metrics for a first site switch and a second site switch, wherein the performance metrics for the first site switch comprise metrics related to one or more host servers associated with the first site switch and reachable via the first site switch and the performance metrics for the second site switch comprise metrics related to one or more host servers associated with the second site switch and reachable via the second site switch;

storing, at the load balancing switch, a plurality of network addresses generated in response to a domain name query; and

ordering, at the load balancing switch, the plurality of network addresses one or more times based upon the stored performance metrics until an ordered list of network addresses is generated that has only one network address at the top of the ordered list;

wherein the ordering comprises ordering the plurality of network addresses using at least one of

round trip time information associated with the first and second site switches, wherein the round trip time information associated with a site switch is indicative of time for exchanging messages between the site switch and a client machine generating the domain name query,

available session capacity associated with network addresses in the plurality of network addresses,

a flashback speed associated with the first site switch and the second site switch, wherein a flashback speed associated with a site switch is indicative of a time required for the site switch to respond to a health check performed by the load balancing switch on the site switch,

geographical locations of the host servers associated with the first and second site switches, or

previous selections of network addresses in the plurality of network addresses as the best network address in response to a domain name query.

13. A system for performing load balancing, the system comprising:

a load balancing switch configured to store performance metrics for a plurality of site switches, each site switch associated with one or more host servers, wherein the one or more host servers associated with a site switch are reachable via the site switch using a virtual address configured at the site switch, the stored performance metrics comprising one or more metrics related to the one or more host servers;

wherein the load balancing switch is further configured to;

store a plurality of network addresses generated in response to a domain name query;

order the plurality of network addresses based upon a first set of performance metrics from the stored performance metrics to generate a first ordered list of network addresses;

determine if a single network address tops the first ordered list generated based upon the first set of performance metrics; and

reorder one or more network addresses from the plurality of network addresses based upon a second set of performance metrics from the stored performance metrics upon determining that a single network address does not top the first ordered list wherein the second set of performance metrics is different from the first set of performance metrics.

14. The system of claim 13 wherein the load balancing switch is configured to:

determine at least one network address from the plurality of network addresses that fails a health check; and

exclude the at least one network address from the first ordered list of network addresses.

15. The system of claim 14 wherein the health check is a Open System Interconnection (OSI) layer 4 or layer 7 health check.

16. The system of claim 14 wherein the health check is a Transmission Control Protocol (TCP) or User Datagram Protocol (UDP) health check.

17. The system of claim 13 wherein the load balancing switch is configured to order or reorder the plurality of network addresses based upon, for each site switch in the plurality of site switches, a number of sessions supported by the site switch and a maximum number of sessions supportable by the site switch.

18. The system of claim 13 wherein the load balancing switch is configured to:

store the performance metrics in a database.

19. The system of claim 13 wherein:

the performance metrics comprise round trip time information for each site switch in the plurality of site switches and a client machine sending the domain name query, wherein the round trip time information for a site switch is indicative of time for exchanging messages between the site switch and the client machine; and

the load balancing switch is configured to order or reorder the plurality of network addresses based upon the round trip time information included in the performance metrics.

20. The system of claim 13 wherein the load balancing switch is configured to order or reorder the plurality of network addresses based upon available session capacity associated with network addresses in the plurality of network addresses.

21. The system of claim 13 wherein the load balancing switch is configured to order or reorder the plurality of network addresses based upon a flashback speed associated with each site switch in the plurality of site switches, wherein a flashback speed for a site switch is indicative of a time required for the site switch to respond to a health check performed by the load balancing server on the site switch.

22. The system of claim 13 wherein the load balancing switch is configured to order or reorder the plurality of network addresses based upon geographical locations of the host servers associated with the plurality of site switches.

23. The system of claim 13 wherein the load balancing switch is configured to order or reorder the plurality of network addresses by placing a network address from the plurality of network addresses of a least recently selected host server for placement at a higher position in the first or second ordered list.

24. A system for performing load balancing, the system comprising:

a load balancing switch adapted to

store performance metrics for a first site switch and a second site switch, wherein the performance metrics for the first site switch comprise metrics related to one or more host servers associated with the first site switch and reachable via the first site switch and the performance metrics for the second site switch comprise metrics related to one or more host servers associated with the second site switch and reachable via the second site switch,

store a plurality of network addresses generated in response to a domain name query, and

order the plurality of network addresses one or more times based upon the stored performance metrics until an ordered list of network addresses is generated that has only one network address at the top of the ordered list;

wherein the load balancing switch is adapted to order the plurality of network addresses using at least one of

round trip time information associated with the first and second site switches, wherein the round trip time information associated with a site switch is indicative of time for exchanging messages between the site switch and a client machine generating the domain name query,

available session capacity associated with network addresses in the plurality of network addresses,

a flashback speed associated with the first site switch and the second site switch, wherein a flashback speed associated with a site switch is indicative of a time required for the site switch to respond to a health check performed by the load balancing switch on the site switch,

geographical locations of the host servers associated with the first and second site switches, or

previous selections of network addresses in the plurality of network addresses as the best network address in response to a domain name query.

25. A computer-readable storage medium storing a plurality of instructions for controlling a processor to perform load balancing, the plurality of instructions comprising:

instructions that cause the processor to store performance metrics regarding a plurality of site switches, each site switch associated with one or more host servers, wherein the one or more host servers associated with a site switch are reachable via the site switch using a virtual address configured at the site switch and wherein the performance metrics regarding a site switch comprise one or more metrics related to one or more host servers associated with the site switch;

instructions that cause the processor to store a plurality of network addresses generated in response to a domain name query;

instructions that cause the processor to order the plurality of network addresses based upon a first set of performance metrics from the stored performance metrics to generate a first ordered list of network addresses;

instructions that cause the processor to determine if a single network address tops the first ordered list generated based upon the first set of performance metrics; and

instructions that cause the processor to reorder one or more network addresses from the plurality of network addresses based upon a second set of performance metrics from the stored performance metrics upon determining that a single network address does not top the first ordered list, wherein the second set of performance metrics is different from the first set of performance metrics.

26. A method performed by a load balancing switch, the method comprising:

performing, at the load balancing switch, a first ordering of a plurality of network addresses based upon a first set of performance metrics for a set of servers, the plurality of network addresses generated in response to a domain name query; and

performing, based upon a result of the first ordering, a second ordering of the plurality of network addresses based upon a second set of performance metrics for the set of servers;

wherein the second set of performance metrics is different from the first set of performance metrics, wherein the first set of performance metrics comprises one of;

round trip time information associated with a plurality of site switches associated with one or more servers from the set of servers, wherein the round trip time information associated with a site switch is indicative of time for exchanging messages between the site switch and a client machine generating the domain name query;

available session capacity associated with network addresses in the plurality of network addresses;

a flashback speed associated with each of the plurality of site switches, wherein a flashback speed associated with a site switch is indicative of a time required for the site switch to respond to a health check performed by the load balancing switch on the site switch;

geographical locations of the set of servers;

or

previous selections of network addresses in the plurality of network addresses as the best network address in response to a domain name query.

27. The method of claim 26 wherein the one or more host servers associated with a site switch are reachable via the site switch using a virtual address configured at the site switch.

28. The method of claim 26 wherein the second ordering is performed upon determining that the first ordering does not yield a single network address as a best candidate from the plurality of network addresses.