Process-based network flow statistical method

The invention provides a process-based network flow statistical method, which comprises the following steps of: capturing a data packet in front of a queue from a data packet queue by a coroutine, performing transmission layer filtering, network layer filtering and application layer filtering, calculating to obtain a value of effective data volume of the data packet, and pushing the value to a data channel, and calculating the value of the valid data volume in the data channel, and continuously obtaining the flow value of the process. The method has the following advantages: (1) the method can be used after being simply opened, and can be quickly used in linux x86 and arm environments at low cost by utilizing a compiled binary execution program; 2, starting from an application layer, flow is filtered based on processes and service names, so that the method is closer to the service purpose, interference factors are reduced, and analysis difficulty is reduced; therefore, the observation difficulty caused by num