Program execution sequence data flow network system and control method thereof

Abstract

The invention provides a network system as well as a control method for the program execution sequence data flow. The network system collocates each arithmetic device an event bus which is respectively connected with a message bus, and simultaneously collocates each arithmetic device a module, so as to enable the mounting and the maintenance to be further simplified. The invention also provides the control method for the program execution sequence data flow, and the control method is characterized in that pure data or internal programs are directly taken as the execution sequence data flow for processing, and the data flow transmission or exchange among the arithmetic devices is controlled by the modules; the message buses are taken as a relay medium, while the event buses are corresponding to the data flow transmission or exchange of the arithmetic devices which are deployed with the event buses, thereby enabling the transmission of the program execution sequence data flow to be more efficient. Partial calculation is shared by the event buses and the message buses, thereby achieving the aim to further alleviate the server operation.

Description

technical field [0001] The present invention relates to a program execution sequence data flow network system and its control method, in particular to a program execution that can make the computing network easy to set up and maintain, can improve the program execution sequence data flow transmission efficiency, and can slow down the network system transmission and computing load Sequential data flow network system and its control method. Background technique [0002] Since the rise of a large number of computing devices and the lowering of the threshold of network technology, the development of computing network (Computing Network) has become more and more huge. Gradually, one-to-one (Point to Point, P2P) or one-to-many (Point to Network) , P2N) connection is no longer sufficient, and the computing speed and computing scale are severely limited by hardware specifications. However, the cost of high-speed high-speed computers is high; a so-called A two-tier (2 Tiers) computi...

AI Technical Summary

Problems solved by technology

[0002] Since the rise of a large number of computing devices and the lowering of the threshold of network technology, the development of computing network (Computing Network) has become more and more huge. Gradually, one-to-one (Point to Point, P2P) or one-to-many (Point to Network) , P2N) connection is no longer sufficient, and the computing speed and computing scale are severely limited by hardware specifications. However, the cost of high-speed high-speed computers is high; a so-called A two-tier (2 Tiers) computing network structure emerged as the times require. In this method, server-side computing devices allocate most of the shared network resources, mainly applications or computing rules, and control the computing time available to each client computing device ( Processing, or, CPU Time), that is, to limit the time for individual computing devices to start and end computing, and to achieve optimal resource allocation through such a network system and computing method, the purpose of which is to reduce any computing that may be delayed due to scheduling Time and wasted computing resources; however, when the scale of the network grows to the order of hundreds, thousands, or even tens of thousands of computing devices instead of dozens of computing devices, the aforementioned resource allocation will inevitably be difficult to achieve in every aspect, and it will inevitably be due to machine performance (Performance ) limit and encounter a bottleneck, although expanding the number of server-side computing devices to reduce the ratio of client computing devices served by a single server-side computing device can improve execution performance, but the increase in the number of server-side computing devices brings, Especially for network administrators, it is not easy for network management and device maintenance, and the workload of monitoring and management; the prior art computing network system schedules the execution or computing time (Processing or Computing Time) of a specific client computing device for a specific program through the server, That is to say, through the means of "scheduling", all the program execution needs become a single execution sequence. When the demand of the client computing device increases, the execution sequence also increases greatly, but the server computing device does not have enough system space. The calculation time will cause the system to be idle (idol) or the execution sequence to be dead. To solve this problem, how to provide an improved network system, especially a network system that can speed up the process and reduce waste as much as possible for the control of the execution sequence. And a relatively more efficient control method based on this network system has become an urgent demand

Images