Test method and system for concurrent user access

Abstract

The invention discloses a test method and system for concurrent user access. The method includes the steps: receiving executive tasks applied for by users by a concurrent scheduling unit; decomposing the tasks into subtasks by the concurrent scheduling unit according to types of the tasks, and placing the subtasks into task scheduling queues; taking out the subtasks from the task scheduling queues by the concurrent scheduling unit to execute the subtask. The executive tasks are decomposed into the subtasks and respectively executed, task execution is scheduled, the users are not directly scheduled, a large number of concurrent user test tasks can be processed under the condition of high concurrent loads of the users, statistical accuracy of response time of the users is ensured, and fairness of user task execution is also ensured.

Description

technical field [0001] The invention relates to the technical field of concurrent access testing, in particular to a testing method and system for user concurrent access. Background technique [0002] When conducting software and hardware performance tests, due to the high concurrent user load characteristics in the performance test model, when the number of users increases to a relatively large number of concurrent accesses, there will be hundreds or even more transaction execution requests at a certain moment. To handle a large number of user tasks, this requires an efficient scheduling method. [0003] At present, the concurrent user scheduling method during stress testing generally uses multi-thread technology to achieve concurrent access by multiple users, and each thread represents a user. However, this method has the following obvious shortcomings: direct scheduling of threads may cause a certain user request to wait for too long to obtain an available database conne...

AI Technical Summary

Problems solved by technology

However, this method has the following obvious shortcomings: direct scheduling of threads may cause a certain user request to wait for too long to obtain an available database connection, and users often cannot accurately count the time from sending to receiving execution of a user request Response time, that is, when the final result is output, it is impossible to correctly count the response time of each user request and the number, proportion, success or failure of transactions, etc.

Images