COMPUTER SYSTEM

• US 20130275626A1
• Filed: 10/25/2010
• Published: 10/17/2013
• Est. Priority Date: 10/25/2010
• Status: Active Grant

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Nishijima, Eiji, Kim, Sungho
• Granted Patent: US 9,942,107 B2
• Time in Patent Office: Days
• US Class Current: 709/248
• CPC Class Codes: G06F 11/1675 Temporal synchronisation or re-synchronisation of redundant processing components
• CPC Class Codes: G06F 11/1687 at event level, e.g. by interrupt or result of polling
• CPC Class Codes: G06F 11/182 based on mutual exchange of the output between redundant processing components
• CPC Class Codes: G06F 11/184 where the redundant components implement processing functionality
• CPC Class Codes: G06F 9/52 Program synchronisation; Mutual exclusion, e.g. by means of semaphores
• CPC Class Codes: H04L 41/50 Network service management, e.g. ensuring proper service fulfilment according to agreements

Abstract

A computer system includes a plurality of computer nodes, each including an external communications unit. An application unit executes processing in accordance with a processing request. A synchronization unit establishes synchronization of the processing between each computer node and other computer nodes. The processing is executed by each computer node, and an inter-node communications unit executes transmission/reception of information between each computer node and the other computer nodes. The synchronization unit transmits the processing request to the other computer nodes via the inter-node communications unit, the processing request being received by the external communications unit. Also, the synchronization unit receives processing requests from the other computer nodes as well via the inter-node communications unit. Based on the number of the computer nodes that have received the same processing request via the external communications units, the synchronization unit selects a processing request that should be executed by the application unit.

Claims

1. A computer system, comprising a plurality of computer nodes,

each of said computer nodes, comprising:

an external communications unit for receiving a processing request;

an application unit for executing said processing in accordance with said processing request;

a synchronization unit for controlling synchronization of said processing between each computer node and the other computer nodes, said processing being executed by said application unit of each computer node; and

an inter-node communications unit for executing transmission/reception of information between each computer node and said other computer nodes, wherein

said synchronization unit of each computer node attaches said computer node'"'"'s identification information to said processing request as transmission-source computer node'"'"'s identification information,

said synchronization unit of each computer node then transmitting said processing request to said other computer nodes via said inter-node communications unit, processing request'"'"'s identification information being attached to said processing request that is received by said external communications unit,

said synchronization unit of each computer node then receiving processing requests from said other computer nodes via said inter-node communications unit, said other computer nodes'"'"' identification information being attached to said processing requests,

said synchronization unit of each computer node then selecting a processing request in accordance with a first predetermined judgment criterion,

said selection of said processing request being made from among said processing requests received from said other computer nodes by said inter-node communications unit, and said processing request received by said external communications unit, said application unit being caused to execute said processing request, said first predetermined judgment criterion being based on the number of said computer nodes that have received a processing request via said external communications units, said same processing request'"'"'s identification information being attached to said processing request, and

said application unit then executing said processing request selected.

2. The computer system according to claim 1, wherein

said first predetermined judgment criterion is whether or not computer nodes, whose number is larger than the half of the number of said plurality of computer nodes, have received said processing request via said external communications units, said same processing request'"'"'s identification information being attached to said processing request.

3. The computer system according to claim 1, wherein

said synchronization unit of each computer node transmits an execution result to said other computer nodes via said inter-node communications unit, said execution result being acquired by said application unit'"'"'s executing said selected processing request,

said synchronization unit of each computer node then receiving execution results from said other computer nodes via said inter-node communications unit,

said synchronization unit of each computer node then selecting an output result in accordance with a second predetermined judgment criterion,

said selection of said output result being made from among said execution results received from said other computer nodes by said inter-node communications unit, and said execution result acquired by said application unit, said second predetermined judgment criterion being based on the number of said computer nodes that have acquired one and the same execution result,

said synchronization unit of each computer node then electing a leader computer-node candidate in accordance with a criterion determined in advance, if said synchronization unit of each computer node has received said selected output result as said execution result from said application unit,

said election of said leader computer-node candidate being made from among said other computer nodes and said self computer node, each of which has transmitted said selected output result as said execution result of said application unit,

said synchronization unit of each computer node then transmitting said leader computer-node candidate'"'"'s identification information to said other computer nodes via said inter-node communications unit,

said synchronization unit of each computer node then receiving leader computer-node candidate'"'"'s identification information from said other computer nodes via said inter-node communications unit, said leader computer-node candidate being elected by said other computer nodes,

said synchronization unit of each computer node then making a judgment as to whether or not said leader computer-node candidate elected by said self computer node should be elected as a leader computer node,

said judgment on said election of said leader computer node being made depending on the number of said computer nodes that have elected, as said leader computer-node candidates, said computer nodes identified by the same identification information, and

said synchronization unit of each computer node then outputting said selected output result, if said leader computer-node candidate elected by said self computer node is elected as said leader computer node, and if said self computer node is elected as said leader computer node.

4. The computer system according to claim 3, wherein

said second predetermined judgment criterion is whether or not computer nodes, whose number is larger than the half of the number of said plurality of computer nodes, have received said one and the same execution result from said application units.

5. The computer system according to claim 3, wherein

said synchronization unit of each computer node supervises an output result, if said leader computer-node candidate elected by said self computer node is elected as said leader computer node, and if said self computer node is not elected as said leader computer node, said output result being outputted by another computer node that is elected as said leader computer node,

said synchronization unit of each computer node then electing another leader computer node, if said output result is different from said output result selected by said self computer node,

said different output result being outputted by another computer node elected as said leader computer node, said election of another leader computer node being made from among said other computer nodes and said self computer node, each of which has selected the same output result as said output result selected by said self computer node.

6. The computer system according to claim 3, wherein

said synchronization unit of each computer node is equipped with process-state information for indicating progress situation of a processing to be executed by said synchronization unit,

said synchronization unit of each computer node further attaching said process-state information to said processing request received by said external communications unit,

said synchronization unit of each computer node then transmitting said processing request to said other computer nodes via said inter-node communications unit,

said synchronization unit of each computer node then receiving processing requests from said other computer nodes via said inter-node communications unit, said other computer nodes'"'"' process-state information being attached to said processing requests,

said synchronization unit of each computer node then making a comparison between said self computer node'"'"'s process-state information, and said other computer nodes'"'"' process-state information received from said other computer nodes, and,

if a delay in said progress of said processing by said synchronization unit of said self computer node is larger than a predetermined criterion,

said synchronization unit of each computer node then transmitting a re-synchronization request to said other computer nodes without executing said processing for selecting said processing request,

said application unit being caused to execute said processing request, said re-synchronization request being made for establishing synchronization between said self computer node and said other computer nodes.

7. The computer system according to claim 6, wherein

said synchronization unit of each computer node manages the number of said computer nodes that have established said synchronization with said self computer node,

said synchronization unit of each computer node transmitting said process-state information to another computer node, if said synchronization unit receives a re-synchronization request from another computer node, and

said synchronization unit of each computer node updating said number of said computer nodes, if said synchronization unit receives a response to said process-state information from another computer node.

8. The computer system according to claim 6, wherein

said synchronization unit of each computer node makes said comparison between said self computer node'"'"'s process-state information, and said other computer nodes'"'"' process-state information received from said other computer nodes, and,

if said delay in said processing by said synchronization unit of said self computer node falls within said predetermined criterion,

said synchronization unit of each computer node executes said processing for selecting said processing request, said application unit being caused to execute said processing request.

9. The computer system according to claim 3, wherein

said synchronization unit of each computer node is equipped with process-state information for indicating progress situation of a processing to be executed by said synchronization unit,

said synchronization unit of each computer node further attaching said process-state information to said execution result of said application unit,

said synchronization unit of each computer node then transmitting said execution result to said other computer nodes via said inter-node communications unit,

said synchronization unit of each computer node then receiving execution results from said other computer nodes via said inter-node communications unit, said other computer nodes'"'"' process-state information being attached to said execution results,

said synchronization unit of each computer node then making a comparison between said self computer node'"'"'s process-state information, and said other computer nodes'"'"' process-state information received from said other computer nodes, and,

if a delay in said progress of said processing by said synchronization unit of said self computer node is larger than a predetermined criterion,

said synchronization unit of each computer node then transmitting a re-synchronization request to said other computer nodes without executing said processing for selecting said output result,

said re-synchronization request being made for establishing synchronization between said self computer node and said other computer nodes.

10. The computer system according to claim 9, wherein

said synchronization unit of each computer node manages the number of said computer nodes that have established said synchronization with said self computer node,

said synchronization unit of each computer node transmitting said process-state information to another computer node, if said synchronization unit receives a re-synchronization request from another computer node, and

said synchronization unit of each computer node updating said number of said computer nodes, if said synchronization unit receives a response to said process-state information from another computer node.

11. The computer system according to claim 9, wherein

said synchronization unit of each computer node makes said comparison between said self computer node'"'"'s process-state information, and said other computer nodes'"'"' process-state information received from said other computer nodes, and,

if said delay in said processing by said synchronization unit of said self computer node falls within said predetermined criterion,

said synchronization unit of each computer node executes said processing for selecting said output result.

12. The computer system according to claim 3, wherein

said synchronization unit of each computer node manages the number of computer nodes that have established synchronization with said self computer node,

said synchronization unit of each computer node transmitting process-state information to another computer node, if said synchronization unit receives a synchronization request from another computer node,

said process-state information being used for indicating progress situation of a processing by said synchronization unit, said synchronization request being made for establishing said synchronization with said self computer node, and

said synchronization unit of each computer node updating said number of said computer nodes, if said synchronization unit receives a response to said process-state information from another computer node.