Self-adaptive thread pool size dynamic adjustment method based on segmentation

Abstract

The invention provides a self-adaptive thread pool size dynamic adjustment method based on segmentation. The method comprises the steps that first, an upper limit T<max> and a lower limit T<min> of the dynamically changing size of a thread pool are configured; second, if the user requested quantity is at a rising stage, adjustment factors are used to increase the size of the thread pool; third, ifthe user requested quantity is at a stable stage, when the user requested quantity is greater than the quantity of work threads in the thread pool, the user requested quantity is used as the goal ofthread pool size adjustment; fourth, if the user requested quantity is at a descending stage, the adjustment factors are used to decrease the size of the thread pool; and fifth, the adjusted size of the thread pool is compared with the upper limit T<max> and the lower limit T<min> respectively, wherein when the size is greater than T<max>, T<max> is used as the size of the thread pool, and when the size is smaller than T<min>, T<min> is used as the size of the thread pool. According to the method, by monitoring the state of the thread pool and adjusting the size of the thread pool, the demandof a user is met to the maximum extent.

Description

technical field [0001] The invention relates to a method for adjusting the size of a thread pool, in particular to a method for dynamically adjusting the size of the thread pool. Background technique [0002] Thread pool has been widely used in Internet technology. By introducing thread pool, threads can be effectively managed, the upper limit of threads can be controlled, and the overhead of creating and destroying threads can be reduced. In practical applications, as the business grows, one system calls more and more other systems. Using a unified thread pool instance in the system can effectively manage threads uniformly and reduce the overhead of creating and destroying threads. [0003] Thread pool technology is the product of reducing the need for thread creation and destruction overhead. It refers to pre-creating a certain number of threads in the server memory, and centrally manages and schedules threads to improve system concurrency. Thread pool technology has natu...

AI Technical Summary

Problems solved by technology

[0006] Thread pool technology is becoming more and more popular in application services. Everything has two sides. Thread pool technology is based on multi-thread technology. Therefore, while overcoming the shortcomings of traditional multi-thread technology, there will still be certain risks in application. ,

[0007] (1) There is no upper or lower limit on the number of threads. Creating and destroying threads too frequently will increase the overhead of the system, which is contrary to the original intention of using the thread pool (reducing the overhead of thread creation and destruction); when requests surge, resources may be exhausted cause system crash

[0008] (2) The number of threads is too large, there are too many threads participating in processing requests in the system, and the increase or decrease of synchronization overhead between multiple threads becomes complicated. At the same time, the system memory overhead for maintaining thread operation increases, which will cause slow system response ;

[0009] (3) The number of threads is too small, and there are no idle threads in the thread pool. When a new request arrives, it needs to wait in the blocking queue, which affects the response speed of the system.

However, the response to tasks or user requests in the thread pool size is ignored, resulting in the failure of the adjustment strategy when using the thread pool on a multi-tasking hardware platform.

Another example is "A Thread Pool Management Method and Device" disclosed in the patent document with application number 201610154360.6. Multiple tasks share the same thread pool, which increases the complexity of thread pool scheduling, and the restrictions on the thread pool are unreasonable, which may lead to The thread pool frequently creates and destroys the thread pool, which increases the resource consumption when the thread pool is used

Images