Data exchange system for static distributed computing architecture and method thereof

Abstract

The invention discloses a cross-computing data exchange system for a static distributed computing architecture. The static distributed computing architecture at least comprises a first computing device and a second computing device. The system comprises a plurality of data executive bodies, a data carrying request collection assembly deployed on each computer and a message transmission bus, wherein a second data executive body on the second computing device sends a second message indicating that data can be read to a first carrying executive body on the first computing device from a message bin of the second data executive body; when the first carrying executive body receives the second message, the finite-state machine of the first carrying executive body changes the state to trigger theoperation assembly to send a first command for executing data reading operation to the first data carrying request collection assembly; and after receiving the first command, the first data transfer request collection component transfers the data in the second output data cache to the first output data cache of the first transfer executive body based on the reading description contained in the first command.

Description

technical field [0001] The present disclosure relates to a data exchange technology, and more specifically, the present disclosure relates to a system and method for implementing data exchange in a static distributed architecture. Background technique [0002] With the popularization of distributed computing, large-scale jobs will deploy different parts of the data to each computing device of different distributed data processing systems for processing through segmentation. In this way, during the processing of a specific job, a The calculation intermediate parameters or results on the computing device will become the input data of the computing task on another computing device, so in order to realize the data synchronization of the intermediate parameters, this will cause the calling overhead of the data transmission between the computing devices. The network communication call is often a bottleneck, and the performance of the network performance communication is not good, ...

AI Technical Summary

Problems solved by technology

[0002] With the popularization of distributed computing, large-scale jobs will deploy different parts of the data to each computing device of different distributed data processing systems for processing through segmentation. In this way, during the processing of a specific job, a The calculation intermediate parameters or results on the computing device will become the input data of the computing task on another computing device, so in order to realize the data synchronization of the intermediate parameters, this will cause the calling overhead of data transmission between computing devices

The network communication call is often a bottleneck, and then the performance of the network performance communication is not good, which will affect the acceleration ratio and scalability of the multi-machine distributed data processing architecture

Images