Parallel processing method and device for complex tasks

Abstract

The invention discloses a parallel processing method and a parallel processing device for complex tasks. The method comprises the following steps that: a request processing unit decomposes a received client request into a plurality of independent subtasks, and sends each subtask to an independent unit for processing; the request processing unit receives a successful processing result or a failed processing result returned by each working unit; when successful processing results of all subtasks are received within a predetermined period of time, the request processing unit combines all successful processing results and sends to a corresponding client; and when the successful processing result of at least one subtask is not returned in a predetermined period of time, the request processing unit interrupts processing of all non-completed subtasks, and performs roll returning operation on subtasks required to be rolled back. Due to the adoption of the parallel processing method and the parallel processing device, the response speed of a single unit request can be increased effectively, response time is shortened, and the resource utilization efficiency of a server side is increased.

Description

technical field [0001] The invention relates to the technical field of computers, in particular to a parallel processing method and device for complex tasks. Background technique [0002] In existing software systems, the client (Client) / server (Server) mode is widely used. In this mode, the client generally submits the request, and then the client waits for the server to return the result; after receiving the client's request, the server performs corresponding processing according to the request, and then returns the processing result to the client. Since a Server can generally connect to multiple Clients, the Server side is usually designed to handle multiple concurrent requests at the same time, and all requests arriving at the same time can be processed in parallel without interfering with each other; and in many key important applications, The client also has certain requirements for the response time of the request. [0003] With the increase in the complexity of var...

AI Technical Summary

Problems solved by technology

[0008] However, in the implementation of the Client / Server mode in the prior art, there are not many implementation methods for the processing mechanism of a single request. Usually, a single request is serially processed and a layered mode is used for processing. Load balancing, using caching technology to improve access speed

And if the synchronous method cannot meet the processing within a limited time, a complex asynchronous processing mode will be used

Therefore, for complex request processing, if the client / server mode adopts the synchronous mode, it will cause the problem of slow processing, resulting in poor user experience

Moreover, if the corresponding processing cannot be completed within the limited time limit, it may cause a decline in the request processing capacity

Although caching technology is used in the existing complex task processing methods to improve the efficiency of data access, it is not very effective for key tasks

Moreover, although the problem of high concurrency and massive data is solved by using distributed technology to linearly expand the system load, the above-mentioned distributed technology will reduce the response speed for a single request

In addition, if the asynchronous processing mode is adopted in the Client / Server mode, it will greatly increase the complexity of the entire system design, which is not conducive to specific implementation and expansion

Images