Load balancing and failover of network services

• US 5,938,732 A
• Filed: 12/09/1996
• Issued: 08/17/1999
• Est. Priority Date: 12/09/1996
• Status: Expired due to Term

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Singhai, Ashish, Radia, Sanjay R., Lim, Swee Boon
• Primary Examiner(s): Rinehart, Mark H.
• Application Number:
  US08/763,289
• Time in Patent Office: 981 Days
• Field of Search: 395/200.59, 395/200.33, 395/200.49, 395/182.03, 395/182.09, 395/72, 395/200.54, 395/200.69, 395/200.56, 709/224, 709/239, 709/226, 709/229
• US Class Current: 709/229
• CPC Class Codes: G06F 11/2025 using centralised failover control functionality
• CPC Class Codes: G06F 11/2035 without idle spare hardware
• CPC Class Codes: G06F 11/2041 with more than one idle spare processing component
• CPC Class Codes: G06F 2209/505 Clust
• CPC Class Codes: G06F 9/5061 Partitioning or combining of resources
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system
• CPC Class Codes: H04L 67/01 Protocols
• 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 67/1017 based on a round robin mechanism
• CPC Class Codes: H04L 67/1034 Reaction to server failures by a load balancer

Abstract

A system and method for many-to-many failover and load balancing establishes a plurality of service groups for providing desired computing services. Each service group comprises a plurality of hosts, and each host within the service group is available to perform the computing services for which that group as a whole is responsible. Each host may belong to a plurality of service groups. Each operational host within a service group transmits periodic messages to each other host within the service group advising of the status of the transmitting host. A leader host evaluates the periodic messages and, where appropriate, dynamicall reassigns responsibility for particular computing services to a host within the group. The reassignment can be due either to failover or load balancing.

Claims

1. A computer program product, comprising:

a computer usable medium having computer readable code embodied therein for causing reassignment of computer services from one host to any of a plurality of other hosts within a service group, the computer program product including;

first computer readable program code devices configured to cause a computer to transmit a message through a service group address setting forth the state of the computer;

second computer readable program code devices configured to cause a computer to receive at least one message through a service group address from at least one other computer; and

third computer readable program code devices configured to respond to at least one message to acquire or release computer services.

2. The computer program product of claim 1 wherein the messages to acquire or release computer services are the result of the failure of another host within the service group.

3. The computer program product of claim 1 wherein the messages to acquire or release computer services are the result of load balancing within the service group.

4. A method for dynamically reassigning computing services to various hosts in a network including the steps of:

establishing a service group for each computing service including a unique Internet Protocol address assigned to each of a plurality of host computers so that the computing service is available as long as one of the plurality of hosts is available within the service group;

assigning responsibility for a computing service to a first host computer within the service group;

transmitting periodically among the host computers of the service group messages through a service group address representative of the state of at least the first host computer;

receiving through a service group address from the remaining host computers within the service group messages representative of the state of such remaining host computers;

evaluating the presence or absence of such messages and their contents; and

reassigning to another host, within the plurality of computers, responsibility for the computing service in response to such evaluation.

5. The method of claim 4 wherein the reassigning step is the result of a lack of a message from one of the hosts.

6. The method of claim 4 wherein the reassigning step is the result of load balancing within the service group.

8. The method of claim 4, further including the step performed by the first host computer of:

assigning responsibility for particular computing services to each of the remaining plurality of host computers.

9. The method of claim 4, further including the steps of:

comparing utilization with a preset maximum utilization setting; and

performing load shedding where utilization of one of the hosts exceeds the preset maximum utilization setting by shifting new connections to other less loaded hosts.

10. The method of claim 4, further including the steps of:

comparing utilization with a preset minimum utilization setting; and

performing load gaining where utilization of one of the hosts drops below the preset minimum utilization by adding the host to a Domain Name Service Zone.

11. The method of claim 4, further including the step of:

using modification of a Domain Name Service Zone to provide load balancing by establishing a configuration parameter that specifies a minimum number of available service addresses that must be present in the Domain Name Service Zone.

12. The method of claim 4, further including the step of:

maintaining the hosts as hot spares.

13. The method of claim 4, further including the step of:

maintaining one of the hosts in a quiesced state.

14. The method of claim 4, further including the step of:

performing failure detection for a total host failure.

15. The method of claim 4, further including the step of:

performing failure detection for a service failure.

16. The method of claim 4, wherein the plurality of hosts are a redundant array of independent servers.

7. Apparatus for providing dynamic reallocation of computing services within a network including a plurality of hosts comprising:

a service group including a plurality of at least three hosts each available to provide a desired computing service including a unique Internet Protocol address assigned to each of the plurality of at least three hosts so that the computing service is available as long as one of the plurality of hosts is available within the service group;

a message portion configured to provide data through a service group address, where the data concerns the state of each of the plurality of hosts; and

a leader portion configured to assign responsibility through a service group address for each of the computing services to any host within the service group in response to the message portion.