The Research on the .Net Framework Thread Pool Based on Multiple Evaluation Criteria

Thread pool is widely used in the implementation of middleware such as web application server and transaction monitor. This paper presents an adaptive threads-pool model fuzzing management framework based on multiple evaluation criteria which can adjust the size of threads pool according to current computing resources state and incoming requests number. Applying this methodology to the .NET thread pool, we find a configuration that increases throughput by 240% compared with the throughput of a poorly chosen configuration. Using model fuzzing reduces the computational requirements to configure the .NET thread pool from machine-years to machine-hours.