Another challenging trend in Internet evolution is the tremendous growth of the infrastructure in every dimension, including bandwidth capacity of links（背景）. Most real-world applications of traffic classification require tools to work online, reporting live information or triggering action according to classification results（目的）. But online traffic classification on modern links requires trade-offs(Limitations) among accuracy, performance, and cost. The practical challenges have led to many published studies with limited evaluation in a simplified environment (currently simply weakens the scenarios) rather than a systematic rigorous analysis of these trade-offs. For example, in order to work online without custom (often prohibitively expensive) hardware (extra hardware support is a sensitive topic), complex DPI classifiers must sacrifice functionality - either analyzing a shorter portion of the payload stream of each traffic flow, or simplifying their pattern matching approaches.
Machine learning techniques require similar compromises (ML method is also necessary to adjust policies to accommodate Online) to Lower or bound The Latency of Classification During Online Execution. The Data Reduction IS GeneRally Implemented by Limiting The Number of packets of A Flow [. 9, 10] (Method 1: reduce the number of packets) Used for extracting Classification features Computational overhead iS Limited by reducing the SET of features [. 11] Used to Classify the traffic, Ideally the using features that CAN BE extracted with. Low Computational Complexity (method 2: reduction feature extraction . complexity) s Some features not Suitable are for Classification Online Because They are Available only AT the End of a flow (method 3: terminating feature stream is no longer used), such as total transferred bytes.
Dainotti A, Pescape A, Claffy K C. Issues and future directions in traffic classification[J]. IEEE network. 2012, 26(1).