Method of managing distributed servers and distributed information processing system using the method

• US 6,256,747 B1
• Filed: 09/24/1998
• Issued: 07/03/2001
• Est. Priority Date: 09/25/1997
• Status: Expired due to Term

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Masuoka, Yoshimasa, Noda, Fumio, Inohara, Shigekazu, Min, Jinghua
• Primary Examiner(s): Iqbal, Nadeem
• Application Number:
  US09/159,784
• Time in Patent Office: 1,013 Days
• Field of Search: 714/4, 714/15, 714/16, 714/25, 714/27, 714/38, 714/39, 714/42, 714/43, 709/201, 709/203, 709/207, 709/205, 709/220, 709/221, 709/229, 709/249
• US Class Current: 714/4.4
• CPC Class Codes: H04L 41/0893 Assignment of logical groups to network elements
• CPC Class Codes: H04L 67/10 in which an application is distributed across nodes in the network software deployment G06F8/60; multiprogramming arrangements G06F9/46
• CPC Class Codes: H04L 69/329 in the application layer [OSI layer 7]

Abstract

In order to effectively make the grasp of operating conditions of a plurality of servers and a cache management in an information system without increasing a time/labor taken by an administrator, the plurality of servers forms a multi-cast hierarchy dynamically reconstructed by virtue of mutual support and the communication of server status, cache directory and validation is performed on the hierarchy. The administrator has not a need of management for cooperation between servers excepting the designation of some other servers for startup thereof. A cache between servers is shared through the exchange of a cache directory and a validation time is reduced, thereby shortening the response time for users.

Claims

1. A distributed server managing method in which one or more computers with server function interconnected by a network are provided, said computers have their unique numbers (ID'"'"'s) and can operate or stop independently, and the destination of communication between the computers is designated by use of said ID so that computers under operation are managed in a manner classified into one or more first groups, the method comprising:

a process with which each of computers belonging to the first groups is informed of the ID'"'"'s of computers forming a group to which that computer belongs;

a process with which when an administrator newly starts a first computer by giving one or more initial ID'"'"'s of computers under operation, said first computer is added into any one of the first groups; and

a process with which in the case where the number of computers in any one of the first groups exceeds a first fixed number, the first groups are reclassified into groups each having computers the number of which is not larger than said first fixed number.

2. A distributed server managing method according to claim 1, wherein each of the first groups has a leader and said adding process includes:

a processing with which a table of computers under operation held by one or more computers corresponding to said initial ID'"'"'s are transmitted from the one or more computers to said first computer;

a processing with which said first computer selects a second computer in said table;

a processing with which said first computer transmits a request for addition of said first computer into the group through said second computer to a third computer, said third computer being a leader of a group to which said second computer belongs; and

a processing with which said third computer adds said first computer into said group.

3. A distributed server managing method according to claim 1, wherein the computers under operation determine a relationship in upper/lower level between the first groups so that the first groups form a hierarchy structure.

4. A distributed server managing method according to claim 1, wherein said reclassifying process includes a processing with which a leader of said group with computers exceeding in number said first fixed number forms a new group, and a processing with which a part of computers belonging to the exceeding group are transferred to said new group.

5. A distributed server managing method according to claim 3, wherein said reclassifying process includes a processing with which a leader of said group with computers exceeding in number said first fixed number transfers, a part of computers forming a group to which said leader belongs, to an adjacent group in said hierarchy structure.

6. A distributed server managing method according to claim 2, wherein said processing for transmitting the request for addition of said first computer into the group includes:

a processing for measuring a speed of communication with a part or all of said computers under operation; and

a processing for transmitting the request for addition to a group to which a computer having a high speed of communication with said first computer belongs.

8. A distributed server managing method according to claim 7, wherein an adjacent group in the hierarchy structure described in claim 3 is used as said second group.

9. A distributed server managing method according to claim 1, wherein a computer belonging to the first groups performs a processing for collecting the ID'"'"'s of computers forming the first groups.

10. A distributed server managing method according to claim 1, wherein each of computers belonging to the first groups has a table for holding the ID'"'"'s of all computers forming a group to which that computer belongs.

11. A distributed server managing method according to claim 1, wherein in the case where a computer belonging to a third group which is one of the first groups transmits one data to a part or all of computers forming the third group, said computer circulates said data successively through the computers forming the third group.

12. A distributed server managing method according to claim 1, wherein in the case where a computer belonging to said third group collects data from one or more computers forming said third group, said computer circulates one message through said one or more computers and each of said one or more computers adds data into said message at the time of circulation.

14. A distributed server managing method according to claim 13, wherein said third server is a computer which belongs to the group described in claim 1 or claim 7.

15. A distributed server managing method according to claim 13, wherein said movement advance notice message is transmitted to said third server through the successive circulation described in claim 11.

18. A distributed server managing method according to claim 17, wherein said second server is a computer which belongs to the group described in claim 1 or claim 7.

19. A distributed server managing method according to claim 17, wherein said revocation advance notice message is transmitted to said second server through the successive circulation described in claim 11.

26. A recording medium in which a computer program for carrying out the distributed server managing method described in claim 1 is recorded.

7. A distributed server managing method in which one or more computers with server function interconnected by a network are provided, said computers have their unique numbers (ID'"'"'s) and can operate or stop independently, and the destination of communication between the computers is designated by use of said ID so that computers under operation are managed in a manner classified into one or more first groups, the method comprising:

a process with which each of computers belonging to the first groups is informed of the ID'"'"'s of computers forming a group to which that computer belongs;

a process with which when an administrator newly starts a first computer by giving one or more initial ID'"'"'s of computers under operation, said first computer is added into any one of the first groups;

a process with which in the case where the number of computers in any one of the first groups exceeds a first fixed number, the first groups are reclassified into groups each having computers the number of which is not larger than said first fixed number;

a process with which in the case where any one of the computers under operation stops or has a trouble or becomes incapable of communication, that computer is deleted from a group to which that computer belongs; and

a process with which in the case where the number of computers in any one of the first groups becomes smaller than a second fixed number, a part or all of computers belonging to the first groups are transferred to a second group.

28. A distributed server managing method according to claim 7, wherein a computer belonging to the first groups performs a processing for collecting the ID'"'"'s of computers forming the first groups.

29. A distributed server managing method according to claim 7, wherein each of computers belonging to the first groups has a table for holding the ID'"'"'s of all computers forming a group to which that computer belongs.

30. A distributed server managing method according to claim 7, wherein in the case where a computer belonging to a third group which is one of the first groups transmits one data to a part or all of computers forming said third group, said computer circulates said data successively through the computers forming said third group.

31. A distributed server managing method according to claim 7, wherein in the case where a computer belonging to said third group collects data from one or more computers forming said third group, said computer circulates one message through said one or more computers and each of said one or more computers adds data into said message at the time of circulation.

32. A recording medium in which a computer program for carrying out the distributed server managing method described in claim 7 is recorded.

13. A distributed server managing method in which there are provided a client which utilizes information and two or more servers each of which holds one or more information and provides information designated by said client, said client and said servers being interconnected by a network, and information held by one server is moved to another server, the method comprising:

a process with which a movement advance notice message giving an advance notice of the movement of first information held by a first server to a second server is transmitted to one or more third servers so that the advance notice is given to said third servers; and

a process with which when receiving said message, said third server stops an operation of moving second information held by said third server to said second server.

16. A distributed server managing method according to claim 13, wherein when receiving said movement advance notice message, said third server stops the operation of moving said second information to said second server in the case where the number of times of movement of information to said second server is not smaller than a first fixed number.

33. A recording medium in which a computer program for carrying out the distributed server managing method described in claim 13 is recorded.

17. A distributed server managing method in which there are provided a client which utilizes information and two or more servers each of which holds one or more information and provides information designated by a request from said client, said client and said servers being interconnected by a network, and data held by the server can be revoked, the method comprising:

a process with which when a first server revokes first information held by said first server, a revocation advance notice message giving an advance notice of the revocation is transmitted from said first server to one or more second servers having a possibility of possessing the copy of said first information; and

a process with which when receiving said revocation advance notice message, said second server stops the revocation of the copy of said first information held by said second server.

20. A distributed server managing method according to claim 17, wherein when said second server revokes second information, said second server recirculates the circulated revocation advance notice message with an advance notice of the revocation added thereinto.

21. A distributed server managing method according to claim 17, wherein said second server holds the number of holds as the estimation of the number of those servers of said first and second servers which have a possibility of possessing said first information or the copy thereof, obtains the number of advance notices of the revocation of said first information described in said revocation advance notice message or the copy thereof, and stops the revocation of said first information held by said second server or the copy thereof if a number resulting from the subtraction of said number of revocation advance notices from said number of holds is not larger than 1.

22. A distributed server managing method in which there are provided a client which utilizes information, a server which holds the copy of one or more information and provides information designated by a request from said client, and two or more external servers each of which provides information to said server or said client, said client, said server and said external servers being interconnected by a network, and the external server responds to validation in which said server or said client inquires of whether or not the copy of information is the latest one, wherein that one of information held by said server to which the reference is made one or more times since the preceding validation is subjected by said server to validation for said external server every predetermined time.

23. A distributed server managing method according to claim 22, wherein a time between 8 hours and 24 hours is used as said predetermined time.

24. A distributed server managing method according to claim 22, wherein instead of the validation every predetermined time, the validation is performed when a time elapsed since the acquisition of said information from the external server falls within a time zone between 8 hours and 24 hours and a load on said server or a load on the network in the vicinity of said server is not larger than a predetermined value.

25. A distributed server managing method according to claim 22, wherein instead of the validation every predetermined time, the validation is performed after the lapse of 8 hours since the acquisition of said information from the external server and before the lapse of 24 hours and in a time zone designated by an administrator.

34. A recording medium in which a computer program for carrying out the distributed server managing method described in claim 22 is recorded.

27. A distributed information processing system provided with a plurality of server computers and at least one client computer which are interconnected by a network, each of said server computers including a server management section for designating a server computer as the designation of communication by use of a unique ID number applied beforehand, classifying designated server computers into one or more groups, adding a newly started server computer into any one of the groups, and reclassifying, server computers in a group the number of server computers of which exceeds a fixed number, into groups each having server computers the number of which is not larger than said fixed number.