Control method for distributed processing system

• US 5,666,484 A
• Filed: 09/12/1989
• Issued: 09/09/1997
• Est. Priority Date: 09/14/1988
• Status: Expired due to Fees

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Suzuki, Yasuo, Mori, Kinji, Kawano, Katsumi, Orimo, Masayuki, Matsuura, Masayoshi, Teranishi, Yuko, Takeuchi, Masuyuki
• Primary Examiner(s): CHUNG, PHUNG M
• Application Number:
  US07/406,234
• Time in Patent Office: 2,919 Days
• Field of Search: 371/16.5, 371/29.1, 371/32, 371/20.6, 364/264.6 MS File, 364/229.2 MS File, 364/239.6 MS File, 364/230 MS File, 364/230.1 MS File , 364/230.2 MS File, 364/230.3 MS File, 364/230.4 MS File, 364/230.5 MS File, 364/230.6 MS File, 370/61, 395/200, 395/275 MS File, 395/575, 395/182.16
• US Class Current: 714/18
• CPC Class Codes: G06F 15/17337 Direct connection machines, e.g. completely connected computers, point to point communication networks coupling between buses G06F13/4004
• CPC Class Codes: G06F 9/5083 Techniques for rebalancing the load in a distributed system

Abstract

A distributed processor system includes a plurality of processors connected via a transmission medium. A control method for a distributed processor system includes the steps of judging, in each processor, its own receiving capability and performing abnormality processing on the basis of the result of the judgment.

Claims

1. A control method for a distributed processor system including a plurality of processors connected via a transmission medium, each processor having data receiving means for receiving data, comprising the steps of:

monitoring, in each processor, the data receiving means therein and judging the receiving capability of said data receiving means;

transmitting to all other processors in said system, from each processor in which a result of said judging indicates an insufficient receiving capability, a message relating to said result of said judging; and

performing, in each of said all other processors, an abnormality processing corresponding to said result of said judging.

2. A control method for a distributed processor system according to claim 1, wherein the data receiving means in each processor includes a receiving buffer for storing data received from said transmission medium, and wherein said judging step comprises the step of judging, in each processor, the receiving capability of the data receiving means of the processor by monitoring the number of data items stored in the receiving buffer of said data receiving means.

3. A control method for a distributed processor system according to claim 2, wherein said abnormality processing comprises the step of setting a flag for indicating an abnormal state when the result of the receiving capability judgment indicates that the receiving means has an insufficient receiving capability.

4. A control method for a distributed processor system according to claim 2, wherein said message includes a data transmission delay requesting message for stopping data transmission for a predetermined time.

7. A control method for a distributed processor system according to claim 4, wherein said abnormality processing comprises the step of repetitively transmitting a first message at predetermined time intervals after transmission of said transmission delay requesting message and transmitting a delay cancel message when the load on said data receiving means falls below a predetermined value.

8. A control method for a distributed processor system according to claim 7, further comprising the steps of temporarily stopping data transmission in a processor which receives a data transmission delay requesting data transmission from the processor which has temporarily stopped data transmission upon receiving a delay cancel message.

9. A control method for a distributed processor system according to claim 8, wherein said abnormality processing further comprises the step of restarting data transmission in the processor when data transmission has been stopped in response to receipt of a data transmission delay requesting message from another processor under the condition that the processor does not receive a second data transmission delay requesting message for a predetermined time or longer.

10. A control method for a distributed processor system according to claim 7, wherein said abnormality processing further comprises the step of restarting data transmission in the processor when data transmission has been stopped in response to receipt of a data transmission delay requesting message from another processor under the condition that the processor does not receive a second data transmission delay requesting message for a predetermined time or longer.

11. A control method for a distributed processor system according to claim 1, wherein said message includes a data transmission delay requesting message for stopping data transmission for a predetermined time.

14. A control method for a distributed processor system according to claim 11, wherein said abnormality processing comprises the step of repetitively transmitting a first message at predetermine time intervals after transmission of said transmission delay requesting message and transmitting a delay cancel message when the load on said data receiving means falls below a predetermined value.

15. A control method for a distributed processor system according to claim 14, further comprising the steps of temporarily stopping data transmission in a processor which receives a data transmission delay requesting message from another processor and restarting data transmission from the processor which has temporarily stopped data transmission upon receiving a delay cancel message.

16. A control method for a distributed processor system according to claim 15, wherein said abnormality processing further comprises the step of restarting data transmission in the processor when data transmission has been stopped in response to receipt of a data transmission delay requesting message from another processor under the condition that the processor does not receive a second data transmission delay requesting message for a predetermined time or longer.

17. A control method for a distributed processor system according to claim 14, wherein said abnormality processing further comprises the step of restarting data transmission in the processor when data transmission has been stopped in response to receipt of a data transmission delay requesting message from another processor under the condition that the processor does not receive a second data transmission delay requesting message for a predetermined time or longer.

5. A control method for a distributed processor system including a plurality of processors connected via a transmission medium, each processor having data receiving means for receiving data, comprising the steps of:

monitoring, in each processor, the data receiving means therein and judging the receiving capability of said data receiving means;

transmitting to other processors, from each processor in which a result of said judging indicates an insufficient receiving capability, a message relating to said result of said judging; and

performing, in each of said other processors, an abnormality processing corresponding to said result of said judging;

wherein the data receiving means in each processor includes a receiving buffer for storing data received from said transmission medium, and wherein said judging step comprises the step of judging, in each processor, the receiving capability of the data receiving means of the processor by monitoring the number of data items stored in the receiving buffer of said data receiving means; and

wherein said message includes a data transmission delay requesting message for stopping data transmission for a predetermined time;

further comprising, in each of said other processors, the steps of;

receiving said data transmission delay requesting message;

storing in a buffer, in each processor receiving said data transmission delay requesting message, data to be transmitted via the transmission medium; and

transmitting said data stored in said buffer upon elapse of a predetermined time period following reception of said data transmission delay requesting message.

6. A control method for a distributed processor system according to claim 5, wherein the length of said predetermined time period is changed according to how frequently the data transmission delay requesting message is received.

12. A control method for a distributed processor system including a plurality of processors connected via a transmission medium, each processor having data receiving means for receiving data, comprising the steps of:

monitoring, in each processor, the data receiving means therein and judging the receiving capability of said data receiving means;

transmitting to other processors, from each processor in which a result of said judging indicates an insufficient receiving capability, a message relating to said result of said judging; and

performing, in each of said other processors, an abnormality processing corresponding to said result of said judging;

wherein said message includes a data transmission delay requesting message for stopping data transmission for a predetermined time;

further comprising, in each of said other processors, the steps of;

receiving said data transmission delay requesting message;

storing in a buffer, in each processor receiving said data transmission delay requesting message, data to be transmitted via the transmission medium; and

transmitting said data stored in said buffer upon elapse of a predetermined time period following reception of said data transmission delay requesting message.

13. A control method for a distributed processor system according to claim 12, wherein the length of said predetermined time period is changed according to how frequently the data transmission delay requesting message is received.

18. A control method for a distributed processor system including a plurality of processors connected via a transmission medium, each processor having data receiving means for receiving data from each of all other processors in said system regardless of whether said data is intended for said each processor, comprising the steps of:

monitoring, in each processor, the data receiving means therein and judging the receiving capability of said data receiving means; and

performing an abnormality processing in each processor in which a result of said judging indicates an insufficient receiving capability;

wherein said abnormality processing comprises the step of stopping data transmission to each of said all other processors for a predetermined period of time when the result of the receiving capability judgment indicates that the data receiving means has an insufficient receiving capability.

19. A control method for a distributed processor system according to claim 18, wherein said abnormality processing further comprises the step of setting a flag for indicating an abnormal state when the result of the receiving capability judgment indicates that the receiving means has an insufficient receiving capability.

20. A control method for a distributed processor system according to claim 18, wherein the data receiving means in each processor includes a receiving buffer for storing data received from said transmission medium, and wherein said judging step comprises the step of judging, in each processor, the receiving capability of the data receiving means of the processor by monitoring the number of data items stored in the receiving buffer of said data receiving means.

21. A transmission method for a distributed processor system according to claim 20, further comprising the step of:

detecting, in the another processor, a collision between said busy signal and the message transmitted by the another processor.

22. A transmission method for a distributed processor system according to claim 21, further comprising the step of:

retransmttting, in the another processor, the message transmitted by the another processor at predetermined time intervals until no collision is detected.

23. In a distributed processor system including a plurality of processors connected via a transmission medium, each processor having a data transmitting and receiving apparatus comprising:

a receiving buffer for storing therein data items received from said transmission medium;

received data monitoring means for monitoring the number of data items stored in said receiving buffer, for comparing the number of data items stored in said receiving buffer with a predetermined number, and for detecting when the number of data items stored in said receiving buffer exceeds said predetermined number;

message sending means for transmitting a transmission delay requesting message to said transmission medium for transmission to each of all other processors in said system when said received data monitoring means detects that the number of data items stored in said receiving buffer exceeds said predetermined number;

data receiving means for receiving data items, including a transmission delay requesting message transmitted by one of the other processors, from said transmission medium and for storing said data items in said receiving buffer;

a transmitting buffer for storing therein data to be transmitted to said transmission medium; and

data transmitting means for transmitting data items stored in said transmitting buffer to the transmission medium, said data transmitting means controlling data item transmission in response to said transmission delay requesting message received by said data receiving means from said transmission medium.

24. A transmission method for a distributed processor system including a plurality of processors connected via a bus type network, said transmission method comprising the steps of:

judging, in each processor, whether the processor is ready to receive a message from said bus type network; and

sending, in each processor which is not ready to receive the message, a busy signal to each of all other processors in said system in response to a result of said judging until the processor is ready to receive the message.

25. A transmission method for a distributed processor system according to claim 24, wherein another processor, which transmits a message and which has detected a collision between the message transmitted by the another processor and said busy signal, retransmits the message transmitted by the another processor at predetermined time intervals until no collision is detected.

26. A transmission method for a distributed processor system according to claim 24, further comprising the step of:

detecting, in each processor which is not ready to receive the message, a message transmitted over said bus type network by another processor;

wherein the sending step is performed when the message transmitted over said bus type network by another processor is detected in the detecting step.