Parallel distributed processing system and method of same

• US 6,098,111 A
• Filed: 03/05/1997
• Issued: 08/01/2000
• Est. Priority Date: 03/05/1996
• Status: Expired due to Term

• Current Assignee: Matsushita Electric Industrial Company Limited (Panasonic Holdings Corporation)
• Original Assignee: Digital Vision Laboratories Corporation
• Inventors: Chiba, Tetsuhiro, Maegawa, Hirotoshi, Saito, Takayuki, Kobayakawa, Yuichi
• Primary Examiner(s): Banankhah, Majid A.
• Assistant Examiner(s): LAO, SUE X
• Application Number:
  US08/812,399
• Time in Patent Office: 1,245 Days
• Field of Search: 709/201, 709/300, 709/303, 709/304, 709/219, 709/217, 709/225, 709/223, 707/10
• US Class Current: 719/310
• CPC Class Codes: G06F 2209/462 Lookup
• CPC Class Codes: G06F 2209/463 Naming
• CPC Class Codes: G06F 9/548 Object oriented; Remote method invocation [RMI] non-remote method invocation G06F9/449
• CPC Class Codes: H04L 67/01 Protocols
• CPC Class Codes: H04L 67/133 Protocols for remote procedure calls [RPC]
• CPC Class Codes: H04L 67/2866 Architectures; Arrangements
• CPC Class Codes: H04L 67/56 Provisioning of proxy services store-and-forward switching systems in data switching networks H04L12/54
• CPC Class Codes: H04L 9/40 Network security protocols

Abstract

A system in which a plurality of network nodes are connected through a network, wherein when a computation object of a first network node calls a function (facility) of a computation object of a second network node, the first network node obtains a facility identification codes (numeral) of the object and the function from a remote reference table and outputs the facility identification codes through the network as a message together with the arguments of the function to the second network node. The second network node uses the input facility identification code as a key to obtain the execution address of the facility from a local reference table and executes that facility.

Claims

1. A parallel distributed processing system wherein one or more computation objects of their facilities executed on a plurality of operation processing nodes connected through a network are mutually called, said parallel distributed processing system comprising:

internal referring means for showing the correspondence between a first identifier identifying the computation object or its facility receiving the call and a second identifier specifying a location of the computation object or its facility receiving the call and specified by a third identifier specifying the module to which the computation object or its facility receiving the call belongs and

external referring means for holding first reference information showing the correspondence among the name or reference of the computation object or its facility receiving the call specified by the computation object or its facility originating the call, the first identifier of the computation object or its facility receiving the call, and a reference of the module to which the computation object or its facility receiving the call belongs and second reference information showing the correspondence among the third identifier of the module, the name of the operation processing node executing the module, and the location information of the operation processing node on the network and specified by the reference of the module,

the computation object or its facility originating the call obtaining, from the name or reference of the computation object or its facility receiving the call obtained by the call using the first reference information of the external referring means, the first identifier and the reference of module to which the computation object or its facility receiving the call belongs, using the second reference information specified using the reference obtained to obtain the third identifier of the module to which the computation object or its facility receiving the call belongs and the location information on the network of the operation processing node executed by the module, and using the obtained first identifier of the computation object or its facility, the obtained third identifier, and the obtained location information of the operation processing node to call the module to which the computation object or its facility belongs,

the module, when receiving the call, using the internal referring means specified using the third identifier to obtain a second identifier corresponding to the first identifier of the computation object or its facility receiving the call and executing the computation object or its facility receiving the call based on the second identifier.

2. A parallel distributed processing system as set forth in claim 1, wherein, when the operation processing node is managed by a predetermined name management system, the correspondence between the name of the operation processing node executing the module shown by the first reference information and the location information of the operation processing node on the network is obtained by converting the name of the operation processing node to the location information of the operation processing node on the network on the basis of said predetermined name management system.

3. A parallel distributed processing system as set forth in claim 2, wherein

a computation space is prescribed which is composed of a set of any of the one or a plurality of computation objects or their facilities and in which uniqueness of names or identifiers of the computation objects or their facilities is required in only the set;

and

the external referring means uses as the name or reference of the receiving the call specified by the of the computation object or its facility originating the call the name or reference in the computation space.

4. A parallel distributed processing system as set forth in claim 3, which prescribes another computation space including at least a computation object or its facility belonging to a different computation space and performs the call in said other computation space.

5. A parallel distributed processing system as set forth in claim 1, wherein, when the module is a process,

said operation processing node comprises:

process allocating means for allocating a program module at a predetermined operation processing node and generating a process at an operation processing node receiving the allocation based on allocation information indicating correspondence between location information of the program module for realizing the process and information of the operation processing node receiving the allocation of the program module and

reference information generating means for generating the reference information of a computation object or its facilities which the program modules for realizing a process mutually refer to based on the allocation information and the reference relationship among the computation objects or their facilities described in the program module in each process.

6. A parallel distributed processing system as set forth in claim 1, wherein

a computation space is prescribed which is composed of a set of any of the one or a plurality of computation objects or their facilities and in which uniqueness of names or identifiers of the computation objects or their facilities is required in only the set;

further comprising

a plurality of reference converting means for converting logical reference information specifying a computation object or its facility in a computation space to either of similar logical reference information and system reference information corresponding to the location information or reference information of a combination of the same; and

conversion control means for recursively inputting the input logical reference information and the logical reference information converted and generated in the reference converting means to any one of the plurality of reference converting means based on the logical reference information and converting the input logical reference information to the system reference information; and

the logical reference information of the process receiving the communication is converted to system reference information by said reference converting means through said conversion control means and communication is performed among processes.

7. A parallel distributed processing system as set forth in claim 1, wherein the call includes a time indicated by a timer used in the module originating the call.

8. A parallel distributed processing system as set forth in claim 1, wherein when a plurality of calls are made from a module originating a call to a module receiving a call, the plurality of calls are made in succession based on a predetermined priority order in the module receiving the call.

9. A parallel distributed processing system as set forth in claim 1, wherein the call includes the time when the facility which is called is to be executed in the module receiving the call.

10. A parallel distributed processing system as set forth in claim 1, wherein the computation object or its facility is any one of a function, class, instance, method, global variable, and file, and the module is a process.

11. A parallel distributed processing system as set forth in claim 1, wherein when a new computation object or facility is registered, the modules were the call of the computation object or facility will be received and originated are analyzed and the reference information is dynamically prepared based on the result of the analysis.

12. An operation processor for executing a process provided with one or more computation objects and connected with other operation processors through a network, said operating processor comprising:

process generating means for generating a process corresponding to a program module,

reference information generating means for generating reference information relating to location information directly specifying on the network a computation object or is facility which said plurality of program modules will refer to together based on a reference relationship among computation objects or their facilities described in the program module, and

reference information holding means for holding said reference information,

said reference information comprising;

internal reference information for showing the correspondence between a first identifier identifying the computation object or its facility receiving the call and a second identifier specifying a location of the computation object or its facility receiving the call and specified by a third identifier specifying the process to which the computation object or its facility receiving the call belongs and

external reference information for holding first reference information showing the correspondence among the name or reference of the computation object or its facility receiving the call specified by the computation object or its facility originating the call, the first identifier of the computation object or its facility receiving the call, and a reference of the process to which the computation object or its facility receiving the call belongs and second reference information showing the correspondence among the third identifier of the process, the name of the operation processing node executing the process, and the location information of other operation processing nodes on the network and specified by the reference of the process,

when originating the call, from the name or reference of the computation object or its facility receiving the call given by the call using the first reference information of the external reference information, the first identifier and the reference of the process to which the computation object or its facility receiving the call belongs, using the second reference information specified using the reference obtained to obtain the third identifier of the process to which the computation object or its facility receiving the call belongs and the location information on the network of the other operation processing nodes executed by the process, and using the obtained first identifier of the computation object or its facility the obtained third identifier, and the obtained location information of the operation processing node to call the process to which the computation object and its facility receiving the call belongs,

when receiving the call, using the internal reference information specified using the third identifier to obtain a second identifier corresponding to the first identifier of the computation object or its facility receiving the call and executing the computation object or its facility receiving the call based on the second identifier.

13. An operation processor for executing a process provided with one or more computation objects and connected with other operation processors through a network, comprising:

a process allocating means for substantively generating a process by instructing the allocation of the program module and a generation of a process corresponding to the program module to a predetermined operation processing node based on allocation information showing the correspondence between location information of a program module for executing the process and information of an operation processing node receiving the allocation of the program module,

reference information generating means for substantially generating reference information by instructing generation, relating to location information directly specifying on the network a computation object or its facility which said plurality of program modules will refer to together, to the process generated corresponding to the program module based on the allocation information and a reference relationship among computation objects or their facilities described in the program module, and

communicating means for instructing the allocation of the process to the operation processing node or instructing the generation of the reference information to the process,

said reference information comprising;

internal reference information for showing the correspondence between a first identifier identifying the computation object or its facility receiving the call and a second identifier specifying a location of the computation object or its facility receiving the call and specified by a third identifier specifying the process to which the computation object or its facility receiving the call belongs and

an external reference information for holding first reference information showing the correspondence among the name or reference of the computation object or its facility receiving the call specified by the computation object or its facility originating the call, the first identifier of the computation object or its facility receiving the call, and a reference of the process to which the computation object or its facility receiving the call belongs and second reference information showing the correspondence among the third identifier of the process, the name of the operation processing node executing the process, and the location information of other operation processing nodes on the network and specified by the reference of the process.

14. A parallel distributed processing method wherein one or more computation objects or their facilities executed on a plurality of operation processing nodes connected through a network are mutually called up, said parallel distribute processing method using:

internal reference information for showing the correspondence between a first identifier identifying the computation object or its facility receiving the call and a second identifier specifying a location of the computation object or its facility receiving the call and specified by a third identifier specifying the module to which the computation object or its facility receiving the call belongs, and

external reference information for holding first reference information showing the correspondence among the name or reference of the computation object or its facility receiving the call specified by the computation object or its facility originating the call, the first identifier of the computation object or its facility receiving the call, and a reference of the module to which the computation object or its facility receiving the call belongs and second reference information showing the correspondence among the third identifier of the module, the name of the operation processing module executing the module, and the location information of the operation processing node on the network and specified by the reference of the module, to resolve references, and

the computation object or its facility originating the call obtaining, from the name or reference of the computation object or its facility receiving the call given by the call using the first reference information of the external reference information, the first identifier and the reference of the module to which the computation object or its facility receiving the call belongs, using the second reference information specified using the reference obtained to obtain the third identifier of the module to which the computation object or its facility receiving the call belongs and the location information on the network of the operation processing node executed by the module, and using the obtained first identifier of the computation object or its facility, the obtained third identifier, and the obtained location information of the operation processing node to call the process to which the computation object and its facility receiving the call belongs, the module, when receiving the call, using the internal reference information specified using the third identifier to obtain a second identifier corresponding to the first identifier of the computation object or its facility receiving the call and executing the computation object or its facility receiving the call based on the second identifier.

15. A parallel distributed processing method as set forth in claim 14, wherein, when the operation processing node is managed by a predetermined name management system, the correspondence between the name of the operation processing node executing the module shown by the first reference information and the location information of the operation processing node on the network is obtained by converting the name of the operation processing node to the location information of the operation processing node on the network on the basis of third predetermined name management system.

16. A parallel distributed processing method as set forth in claim 15, wherein

a computation space is prescribed which is composed of a set of any of the one or a plurality of computation objects or their facilities and in which uniqueness of names or identifiers of the computation objects or their facilities is required in only the set;

and

the external reference information uses as the name or reference of the computation object or its facility receiving the call specified by the computation object or its facility originating the call name or reference of the computation object or its facility in the computation space.

17. A parallel distributed processing method as set forth in claim 14, which prescribes another computation space including at least a computation object or its facility belonging to a different computation space and performs the call in said other computation space.

18. A parallel distributed processing method as set forth in claim 14, comprising, when the module is a process,

allocating a program module at a predetermined operation processing node and generating a process at an operation processing node receiving the allocation based on allocation information indicating correspondence between location information of the program module for realizing the process and information of the operation processing node receiving the allocation of the program module, and

generating the reference information of a computation object or its facility which the program modules for realizing a process mutually refer to based on the allocation information and the reference relationship among the computation objects or their facilities described in the program module in each process.

19. A parallel distributed processing method as set forth in claim 14, further comprising:

prescribing a computation space which is composed of a set of any of the one or a plurality of computation objects or their facilities and in which uniqueness of names or identifiers of the computation objects or their facilities is required in only the set;

determining a method of reference conversion based on the content of logical reference information specifying a computation object or its facility receiving a communication of a module in a computation space;

using the method of reference conversion determined to convert to the logical reference information receiving the communication to either of another logical reference information and system reference information directly specifying an object on the computer network;

determining the method of reference conversion of the logical reference information in the reference information resulting from the conversion based on said logical reference information again;

using the method of reference conversiondetermined to again perform conversion;

repeatedly determining the method of reference conversion for the logical reference information and performing conversion based on it until all elements of the reference information resulting from the conversion become system reference information and communicating among modules based on said reference information when all of the elements become system reference information; and

performing the distributed processing with the modules substantively linked.

20. A parallel distributed processing method as set forth in claim 14, wherein the call includes a time indicated by a timer used in the module originating the call.

21. A parallel distributed processing method as set forth in claim 14, wherein when a plurality of calls are made from a module originating a call to a module receiving a call, the plurality of calls are made in succession based on predetermined priority order in the module receiving the call.

22. A parallel distributed processing method as set forth in claim 14, wherein the call includes a time when the facility which is called is to be executed in the module receiving the call.

23. A parallel distributed processing method as set forth in claim 14, wherein:

the computation object or its facility is any one of a function, class, instance, method, global variable, and file, and the module is a process.

24. A parallel distributed processing method as set forth in claim 14, wherein when a new computation object or facility is registered, the modules where the call of the computation object or facility will be received and originated are analyzed and the reference information is dynamically prepared based on the result of the analysis.