Method and system for error recovery for cascaded servers

• US 5,396,613 A
• Filed: 11/05/1992
• Issued: 03/07/1995
• Est. Priority Date: 11/05/1992
• Status: Expired due to Term

• Current Assignee: University of Utah Research Foundation (State of Utah)
• Original Assignee: University of Utah Research Foundation (State of Utah)
• Inventors: Hollaar, Lee A.
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): Elmore, Stephen C.
• Application Number:
  US07/980,165
• Time in Patent Office: 852 Days
• Field of Search: 395/575, 371/62, 364/945, 364/266.5
• US Class Current: 714/4.1
• CPC Class Codes: G06F 11/1402 Saving, restoring, recovering or retrying

Abstract

An improved method for error recovery in client-server distributed processing systems using cascaded servers. The method adds an additional form of acknowledgement message from a server to its predecessor. This message is sent when the server has completed all its processing and has received an ACK message from its successor, indicating that its result has been successfully received by the successor. The predecessor server retains a copy of its results until it receives the done message, at which time the copy of the results is discarded. If a done message is not received by the time a timer has expired, indicating that a problem has occurred with a server or its communications, the predecessor server resends its stored results.

Claims

1. A method for error recovery in a client-server distributed processing system using cascaded servers, said distributed processing system comprising:

an intermediate server; and

a successor server of said intermediate server, not in a processing loop with said intermediate server, the method comprising;

said intermediate server saving a copy of a request message sent to said successor server; and

if no done message has been received in a predetermined period of time, said intermediate server resending said request message to said successor server.

2. A method as in claim 1, wherein said request message specifies at least one task to be performed by said successor server, and said done message is sent by said successor server after said successor server has completed said at least one task.

3. A method as in claim 1, wherein said done message is sent by said successor server after said successor server has received an acknowledge message in response to a request message.

4. A method as in claim 1, wherein said request message includes a unique identifier and said done message includes a unique identifier corresponding to said request message'"'"'s unique identifier.

5. A method as in claim 4, wherein said request message'"'"'s unique identifier is a timestamp determined when said request message is first sent to said successor server.

6. A method as in claim 4, wherein said request message'"'"'s unique identifier is a unique identifier assigned when a client first makes a request.

7. A method as in claim 1, wherein said predetermined period of time is indicated by a timer interrupt.

8. A method as in claim 1, wherein said predetermined period of time is indicated by completion of a synchronous wait operation.

9. A method as in claim 1, wherein said predetermined period of time is selected based on an expected completion time of said successor server.

10. A method as in claim 9, wherein said expected completion time of said successor server is contained in an acknowledgement message sent to said intermediate server from said successor server in response to said request message.

11. A method as in claim 10, wherein said successor server sends a still-processing message to said intermediate server if said successor server determines that said expected completion time is different than said successor server'"'"'s new estimate of said completion time.

12. A method as in claim 1, wherein said resent request message includes a flag to indicate that it is a resent request message.

13. A method as in claim 12, wherein said successor server sends a negative acknowledge message to said intermediate server if said successor server receives a request message whose flag indicates it is a resent message, and said request message has already been processed.

14. A method for error recovery in a client-server distributed processing system using cascaded servers, said distributed processing system comprising:

an intermediate server;

a successor server of said intermediate server, not in a processing loop with said intermediate server; and

a subsequent server of said intermediate server, the method comprising;

said intermediate server saving a copy of a request message sent to said successor server; and

if no done message has been received in a predetermined period of time, said intermediate server resending said request message to said successor server.

15. A method as in claim 14, wherein said request message specifies at least one task to be performed by said successor server, and said done message is sent by said subsequent server after said successor server has completed said at least one task.

16. A method as in claim 14, wherein said done message is sent by said subsequent server after said subsequent server has received an acknowledge message in response to a request message.

17. A method as in claim 14, wherein said request message includes a unique identifier and said done message includes a unique identifier corresponding to said request message'"'"'s unique identifier.

18. A method as in claim 17, wherein said request message'"'"'s unique identifier is a timestamp determined when said request message is first sent to said successor server.

19. A method as in claim 17, wherein said request message'"'"'s unique identifier is a unique identifier assigned when a client first makes a request.

20. A method as in claim 14, wherein said predetermined period of time is indicated by a timer interrupt.

21. A method as in claim 14, wherein said predetermined period of time is indicated by completion of a synchronous wait operation.

22. A method as in claim 14, wherein said predetermined period of time is selected based on an expected completion time of said successor server.

23. A method as in claim 22, wherein said expected completion time of said successor server is contained in an acknowledgement message sent to said intermediate server from said successor server in response to said request message.

24. A method as in claim 23, wherein said successor server sends a still-processing message to said intermediate server if said successor server determines that said expected completion time is different than said successor server'"'"'s new estimate of said completion time.

25. A method as in claim 14, wherein said resent request message includes a flag to indicate that it is a resent request message.

26. A method as in claim 25, wherein said successor server sends a negative acknowledge message to said intermediate server if said successor server receives a request message whose flag indicates it is a resent message, and said request message has already been processed.

27. A computer system providing error recovery for a client-server distributed processing system having a cascade of servers, comprising:

an intermediate server computer having;

a program implementing a server function,

a memory means for storing information, and

a timing means for determining whether a predetermined period of time has elapsed;

a successor server computer, not in a processing loop with said intermediate server computer said successor server computer being a successor of said intermediate server computer in said cascade of servers and not in a processing loop with said intermediate server computer; and

a data communications system, comprising means for sending messages from said intermediate server computer to said successor server computer and for sending messages from said successor server computer to said intermediate server computer,

wherein said intermediate server computer sends a request message to said successor server computer using said data communications system;

and wherein said intermediate server computer saves said request message in said memory means of said intermediate server computer;

and wherein said intermediate server computer starts said timing means when said request message is sent to said successor server computer;

and wherein said intermediate server computer resends said request message to said successor server computer if said done message has not been received before said timing means has determined whether said predetermined period of time has elapsed.

28. A system as in claim 27, wherein said request message specifies at least one task to be performed by said successor server computer, and said done message is sent by said successor server computer after said successor server computer has completed said at least one task.

29. A computer system providing error recovery for a client-server distributed processing system having a cascade of servers, comprising:

an intermediate server computer having;

a program implementing a server function,

a memory means for storing information, and

a timing means for determining whether a predetermined period of time has elapsed;

a successor server computer, not in a processing loop with said intermediate server computer said successor server computer being a successor of said intermediate server computer in said cascade of servers and not in a processing loop with said intermediate server computer;

a subsequent server computer having a program implementing a server function, said subsequent server computer being a subsequent server in said cascade of servers to said intermediate server computer; and

a data communications system, comprising means for sending messages from said intermediate server computer to said successor server computer and further comprising means for sending messages from said subsequent server computer to said intermediate server computer,

wherein said intermediate server computer sends a request message to said successor server computer using said data communications system;

and wherein said intermediate server computer saves said request message in said memory means of said intermediate server computer;

and wherein said intermediate server computer starts said timing means when said request message is sent to said successor server computer;

and wherein said intermediate server computer resends said request message to said successor server computer if said done message has not been received before said timing means has determined whether said predetermined period of time has elapsed.

30. A system as in claim 29, wherein said request message specifies at least one task to be performed by said successor server computer, and said done message is sent by said subsequent server computer after said successor server computer has completed said at least one task.