US 4,733,347 A · Method and apparatus for synchronized concurrent processes

Abstract

A synchronized concurrent process system incorporating a data driven control with a time attribute being assigned to the shared data. The term "data driven control" means a process operation sequence control system which satisfies the following three conditions. In condition 1, a process is allowed to initiate its operation when all the data items to be referred to by the process are generated by other processes; in condition 2, during a time period from when a processing initiates its operation to when the processing terminates its operation, the other processes are prohibited to update any data item to be referred to by the process, and in condition 3, only when a process terminates its operation, the other processes are allowed to refer to all the data items generated by the process. However, the data driven control alone is not satisfactory. Recognizing the existence of shared data, a time attribute (interior clock) is provided to discriminate the same data before and after updating. This provision has advantages: (i) The procedures for the synchronous control and the mutual exclusive control need not be inserted in an intermediate location of a program, (ii) all processes can operate concurrently (The wait state is not necessary); and (III) a computer system can be provided with the deterministic characteristic.

Claims5

Dependent claims

A process execution control method for use in an information processing system which executes a plurality of processes each referring to at least one data item, to produce, as a result of execution, at least one data item, comprising:

a first step of storing, for each of said processes, an identifier of a data item to be referred to by each process and a time condition representative of a valid time attribute to be assigned to said data item;
a second step of storing, for each of said processes, time attribute information representative of a time point at which each process is to be executed;
a third step of storing a data item produced by executing a given one of said processes and said time attribute information of said second step associated with said executed process;
a fourth step of referring to time attribute information of each of the data items which have been already produced and selecting out of said plurality of processes one process in which every data item to be referred to by said one process satisfies said time condition previously defined at said first step;
and
a fifth step of executing said selected process and updating said time attribute information corresponding to said executed process, so that said time attribute information of said executed process indicates an updated time point;
whereby data items having an identical identifier are given different time attribute informations depending upon time points of execution of a process by which execution of said data items have been produced, and a data item identified by said time attribute information is referred to by each process.

4 more elements

An information processing system which executes a plurality of processes each referring to at least one data item and produces new data as a result of the execution, comprising:

first means for previously storing, for each process, a name of each data item to be referred to by said each process and a time difference necessary for executing each process;
second means for storing, for each process, time attribute information representative of a time point at which said each process is to be executed;
third means for executing a given one of said processes to produce a data item and storing the produced data item and said time attribute information produced by said executed process;
and
fourth means for selecting out of said plurality of processes one process which s one that all of said data items to be referred to by said one process have been already produced, on the basis of said time condition, stored in said first means, of each data item, said time attribute information, stored in said second means, of each process and said time attribute information of each data item which has been already produced by said third means, and for issuing to said third means, when said third means has completed the execution of a given one process, a request for executing said selected process, together with time attribute information to be assigned to a data time to be produced by said selected process, said fourth means updating said time attribute information stored in said second means so as for said time attribute information to indicate a value representing an updated time point with respect to said selected process to which said request for execution is issued;
whereby data items having an identical name are given different time attribute informations depending upon time points of execution of a process by which execution said data items have been produced and a data item identified by said time attribute information is referred to by each process.

3 more elements

Dependencies are inferred from the claim text.

Method and apparatus for synchronized concurrent processes

Life of this patent

Method and apparatus for synchronized concurrent processes

Life of this patent

Abstract

Claims5