Efficient maintenance of a distributed system membership view

• US 9,609,055 B2
• Filed: 08/17/2016
• Issued: 03/28/2017
• Est. Priority Date: 02/17/2009
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Garcs-Erice, Luis
• Primary Examiner(s): Keehn, Richard G
• Application Number:
  US15/239,326
• Publication Number:
  US 20160359967A1
• Time in Patent Office: 223 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 15/163 Interprocessor communication
• CPC Class Codes: G06F 15/173 using an interconnection network, e.g. matrix, shuffle, pyramid, star, snowflake
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: H04L 67/1046 Joining mechanisms
• CPC Class Codes: H04L 67/1053 with pre-configuration of logical or physical connections with a determined number of other peers
• CPC Class Codes: H04L 67/51 Discovery or management thereof, e.g. service location protocol [SLP] or web services
• CPC Class Codes: H04W 4/08 User group management
• CPC Class Codes: H04W 8/186 Processing of subscriber group data

Abstract

A method, system and computer program product are disclosed for maintaining a membership view of a distributed computing system comprising a multitude of members, wherein each of said members sends out a message indicating that said each of the members is available in the system. In one embodiment, the method comprises maintaining a first list of at least some of the members in the distributed system; and over a period of time, when one of the members sends out the message indicating that said one of the members is available in the distributed system, adding said one of the members to a second list. At defined times, the first list is replaced with the second list. In an embodiment, the method further comprises at said defined times, starting a new, empty second list.

Claims

1. A method of removing inactive members from a list of members in a distributed computing system, wherein active members of the computing system generate messages indicating that the active members are available in the computing system, and new members are added to the computing system, the method comprising:

concurrently maintaining first and second lists for members of the computing system, the first list comprising a working list including current members of the computing system, and the second list comprising a listening list, including

adding to the listening list, any of the current members of the computing system that, during a given time period, generate the messages indicating that said any of the current members are available in the computing system, and

adding to the listening list and to the working list, any new members added to the computing system that, during the given time period, generate the messages indicating that said any new members are available in the computing system; and

at a specified time, replacing the working list with the listening to form a new working list, and starting a new, empty listening list, thereby removing from the working list any of said current members of the computing system that did not generate, during the given time period, the messages indicating that said current members are available in the computing system.

2. The method according to claim 1, wherein each of the members in the computing system periodically sends the messages indicating that said each member is available in the computing system.

3. The method according to claim 1, wherein over said given time period, the listening list becomes an updated, complete list of all the members in the distributed computing system that, during said given time period, generated the messages indicating that the members are available in the distributed system.

4. The method according to claim 1, wherein each member that is available in the computing system sends out the message that indicates said each member is available in the system, at a defined time period T_o.

5. The message according to claim 4, wherein said given time period is greater than or equal to T_o.

6. The method according to claim 5, wherein said given time period is greater than T_o.

7. The method according to claim 6, wherein said given time period is three times as long as T_o.

8. The method according to claim 7, wherein said specified time is at the end of the given time period.

9. The method according to claim 1, wherein the replacing the working list includes replacing the working list with the listening list without verifying the availability in the computing system of any of the members on the listening list.

10. The method according to claim 1, wherein:

the working list identifies the location and identity of each of the members on the working list; and

the second list identifies the location and identity of each of the members on the second list.

11. A membership view maintenance system for removing inactive members from a list of members in a distributed computing system, wherein active members of the computing system generate messages indicating that the active members are available in the computing system, and new members are added to the computing system, the membership view maintenance system comprising:

a computer system comprising a memory for storing data, and one or more hardware processor units connected to the memory for transmitting data to and receiving data from the memory, the one or more hardware processor units configured for;

concurrently maintaining first and second lists for members of the computing system, the first list comprising a working list including current members of the computing system, and the second list comprising a listening list, including

adding to the listening list, any of the current members of the computing system that, during a given time period, generate the messages indicating that said any of the current members are available in the computing system, and

adding to the listening list and to the working list, any new members added to the computing system that, during the given time period, generate the messages indicating that said any new members are available in the computing system; and

at a specified time, replacing the working list with the listening to form a new working list, and starting a new, empty listening list, thereby removing from the working list any of said current members of the computing system that did not generate, during the given time period, the messages indicating that said current members are available in the computing system.

12. The membership view maintenance system according to claim 11, wherein each of the members in the computing system periodically sends the messages indicating that said each member is available in the computing system.

13. The membership view maintenance system according to claim 11, wherein over said given time period, the listening list becomes an updated, complete list of all the members in the distributed computing system that, during said given time period, generated the messages indicating that the members are available in the distributed system.

14. The membership view maintenance system according to claim 11, wherein each member that is available in the computing system sends out the message that indicates said each member is available in the system, at a defined time period T_o.

15. The membership view maintenance system according to claim 11, wherein said given time period is greater than or equal to T_o.

16. An article of manufacture comprising:

at least one computer usable device having computer readable program code logic to execute a machine instruction in a processing member for maintaining a membership view of a distributed computing system by removing inactive members from a list of members in the distributed computing system, wherein active members of the computing system generate messages indicating that the active members are available in the computing system, and new members are added to the computing system, the computer readable program code logic, when executing, performing the following;

concurrently maintaining first and second lists for members of the computing system, the first list comprising a working list including current members of the computing system, and the second list comprising a listening list, including

adding to the listening list, any of the current members of the computing system that, during a given time period, generate the messages indicating that said any of the current members are available in the computing system, and

adding to the listening list and to the working list, any new members added to the computing system that, during the given time period, generate the messages indicating that said any new members are available in the computing system; and

at a specified time, replacing the working list with the listening to form a new working list, and starting a new, empty listening list, thereby removing from the working list any of said current members of the computing system that did not generate, during the given time period, the messages indicating that said current members are available in the computing system.

17. The article of manufacture according to claim 16, wherein each of the members in the computing system periodically sends the messages indicating that said each member is available in the computing system.

18. The article of manufacture according to claim 16, wherein over said given time period, the listening list becomes an updated, complete list of all the members in the distributed computing system that, during said given time period, generated the messages indicating that the members are available in the distributed system.

19. The article of manufacture according to claim 16, wherein each member that is available in the computing system sends out the message that indicates said each member is available in the system, at a defined time period T_o.

20. The article of manufacture according to claim 19, wherein said given time period is greater than or equal to T_o.