US 5,313,632 AGranted

Method for processing data in a distributed processing system

AssigneeHitachi, Ltd.

InventorsToshihiko Sekizawa, Kinji Mori, Masayuki Orimo, Yasuo Suzuki, Katsumi Kawano, Minoru Koizumi, Kozo Nakai, Hirokazu Kasashima

Life of this patent

Estimating the term…

patentgrep

Loading patent…

Abstract

Programs in a plurality of processors connected through a common network are started when a plurality of data necessary for starting the program are received from the network. When the processor receives data from an external equipment, it adds the event number to the data and sends it to the network, and each processor determines whether all data necessary for starting the program in itself have been received and checks the relation among the received data by comparison with the event numbers of the received data.

Claims8

Dependent claims

A method for determining whether to start processing a predetermined number of data sets by one of a plurality of programs in a distributed processing system having a plurality of processors interconnected through a common transmission line, comprising the steps of:

(a) adding, in each processor, an initial label which indicates a source of data to a data set generated by said each processor, and sending said generated data set together with said initial label to the said common transmission line;
(b) adding, in each processor which receives said generated data set and said initial label from said common transmission line, another label related to said initial label received with said generated data set to an output data set produced by executing a program for processing said generated data set and sending said output data set together with said another label to said common transmission line;
(c) receiving, in each processor having a program or processing a predetermined number of data sets, said generated and output data sets and said initial and another labels from said common transmission lines;
and
(d) comparing, in said each processor having said program for processing a predetermined number of data sets, said initial and another labels received with said generated and output data sets by said each processor to each other and starting said program for processing a predetermined number of data sets when said comparison indicates that said initial and another labels are related to each other.

2 more elements

A method of determining whether to start processing a predetermined number of data by one of a plurality of programs in a distributed processing system having a plurality of processors connected through a common network and enhancing execution of said programs, comprising the steps of:

(a) adding, in each processor, an initial event number, which indicates a source of data to external data received from an external device in said each processor connected to the external device;
(b) sending, by said each processor, said external data together with said initial event number to the common network;
(c) adding, in each processor which receives said external data and said initial event number from the common network, another event number related to said initial event number included in said received external data to data produced by processing the received external data in said each processor which receives said external data and said initial event number, and sending said produced data together with said another event number to the common network;
(d) receiving, in each processor having a program for processing a predetermined number of data, said external and produced data and said initial and another event numbers from the common network;
(e) identifying, in said each processor having said program for processing a predetermined number of data, event numbers included in said external and produced data and said initial and another event numbers received from the common network that are related to each other;
and
(f) starting, in said each processor having said program for processing a predetermined number of data, said program for processing said predetermined number of data having event numbers identified as being related to each other.

4 more elements

Dependencies are inferred from the claim text.