A resource allocation method and device based on the task packet

Abstract

The disclosed resource allocation method based on task grouping comprises: grouping received task to obtain different task groups and determine relative expectation on resource, then allocating resource according to their expectation to balance the chance for every task group. It also discloses the corresponding devices as task grouping and management modules.

Description

technical field [0001] The invention relates to resource allocation technology, in particular to a resource allocation method and device based on task grouping. Background technique [0002] In the server program based on the client / server (C / S) model, rationally allocating resources on the server side can enable the server program to better serve the client. Server-side resources include: threads, processors, memory, bandwidth, and more. [0003] A thread is a single sequential flow of control within a process. Compared with creating a process, creating a thread will consume much less system resources. For those applications with a particularly large number of concurrent processes, using threads will achieve better performance than using processes. [0004] In order to improve the utilization rate of the processor, multi-threading technology has appeared. Multithreading technology solves the problem of concurrent execution of multiple threads in the processor unit, which...

AI Technical Summary

Problems solved by technology

[0021] If the thread resources in SCPa are scheduled according to the existing thread resource scheduling method, when the number of calls per second is relatively large, it is possible that the thread resources of SCPa are occupied by IDP and ACR, and the ERB in this process The task cannot be processed by the SCPa thread in time, resulting in session timeout and increased call loss rate

[0022] As can be seen from the above description, since heavyweight tasks occupy all thread resources, it will have adverse effects on server applications and client programs.

For the server program, during the time period when all thread resources are occupied by heavyweight tasks, other tasks in the task queue cannot be scheduled and processed, so that more and more tasks are accumulated in the task queue, and the server program occupies more resources. The system memory is also getting larger and larger, which may eventually lead to task queue congestion, and even cause the server application to crash

For the client program, the response to the client's request will not be timely, and when a task of the client is composed of multiple tasks of different heavyweights, since all thread resources of the server are occupied by heavyweight tasks, other orders of magnitude The task cannot be responded to, resulting in the failure of the client's business

[0023] In short, in the existing thread pool scheduling method, some tasks occupy all resources, so that other tasks cannot obtain the service of thread resources.

Images