Thread application instance

One . Wait timeout mode

scene ： Wait for a period of time when calling a method , If this method can get results within a given time , Will return immediately , conversely , If timeout, return the default result .

Two . Database connection pool

Definition of connection pool ： Constructor initializes the maximum limit of the connection , Maintain connections through a two-way queue , The caller needs to call fetchConnection(long) Method to specify how many milliseconds to timeout to get a connection , When the connection is used , Need to call releaseConnection(Connection) Method to put the connection back into the connection pool .

3、 ... and . Thread pool technology

3.1 Role of thread pool

         face Thousands of people Service delivery hand over Into the clothing service device when , If you still use one service One Line The way of Cheng , Then it will gen Build tens of thousands remember Of Line cheng , this It's not a good choice , Because this will make the operating system switch thread context frequently , Increase the load of the system for no reason , The creation and extinction of threads require system resources , It also undoubtedly wastes system resources .

Thread pool solves this problem well , It creates several threads in advance , And it can't be directly controlled by users to create threads

 3.2 Advantages of thread pool

• Reduce resource consumption
• Improve response time
• Improve the manageability of threads

3.3 Several parameters in the thread pool

1: Number of core threads (corePoolSize)
The design of the number of core threads needs to be determined according to the processing time of the task and the number of tasks generated per second , for example : Performing a task requires 0.1 second , The system is... Percent 80 Every second of the time will produce 100 A mission , So if you want to be in 1 Finish this in seconds 100 A mission , Need 10 Threads , At this point, we can design the number of core threads as 10; Of course, the actual situation cannot be so average , So we usually follow 8020 Principle design is enough , Both according to percent 80 Design the number of core threads in the case of , The remaining percent 20 The maximum number of threads can be used to process ;
2: Task queue length (workQueue)
The length of task queue is generally designed as : Number of core threads / Execution time of single task *2 that will do ; For example, in the above scene , The number of core threads is designed as 10, The execution time of a single task is 0.1 second , The queue length can be designed as 200;
3: Maximum number of threads (maximumPoolSize)
The design of the maximum number of threads needs to refer to the conditions of the number of core threads , You also need to refer to the maximum number of tasks generated by the system per second : for example : In the above environment , If the maximum task generated by the system per second is 1000 individual , that , Maximum number of threads =( Maximum number of tasks - Task queue length )* Execution time of single task ; Both : Maximum number of threads =(1000-200)*0.1=80 individual ;
4: Maximum free time (keepAliveTime)
The design of this parameter completely refers to the system operating environment and hardware pressure setting , There is no fixed reference value , The user can set a reasonable value according to experience and the time interval of the task generated by the system ;

3.4 Thread pool workflow

 

3.5Executor and Executors The difference between

Executor Interface objects can perform our threading tasks ;

Executors Different methods of tool classes create different Thread pool , To meet the needs of the business .

ExecutorService Interface inherited Executor Interface and extended , More ways , We can get the execution status of the task and the return value of the task .

 

3.6 Do you know how to create a thread pool ？

ThreadPoolExecutor