Network Traffic Identification Using Machine Leaning and DPI

Abstract

Exact system traffic distinguishing proof is a significant reason for arrange traffic checking and information examination, and is the way to improve the nature of client administration. In this paper, through the examination of two system traffic recognizable proof techniques dependent on Machine Learning (ML) and profound parcel assessment, a system traffic distinguishing proof strategy dependent on ML and profound bundle review is proposed. This technique utilizes profound parcel review innovation to distinguish most system traffic, decreases the remaining burden that should be recognized by ML strategy and profound bundle examination can distinguish explicit application traffic, and improves the exactness of identification. ML strategy is utilized to help with distinguishing system traffic with encryption and obscure highlights, which compensates for the weakness of profound bundle review that can't recognize new applications and encoded traffic. To reduce the work load of Machine Learning we are using Deep Packet Inspection (DPI) Technology. Analysis shows that this strategy can improve the ID pace of system traffic.