Traffic classification and its applications to modern networks

The observation of network traffic is at the core of many fundamental networkoperationandmaintenanceactivities, such as the enforcement of quality of service guarantees or traffic engineering, and of security issues such as intrusion detection and prevention. However, the rapid introduction of new categories of applications such as network games and peer-to-peer communication, the increasing presence of malicious traffic, and the widespread use of encryption techniques, make themeasurement, analysis and classification of Internet traffic a challenging task. The research community has devoted a great effort to the study of new traffic characterization and classification mechanisms, with the specific intent of surpassing the limitations that traditional approaches based on port and payload analysis face. The purpose of this special issue is to collect high-quality contributions in this fertile area of research, with the specific intent to present novel classification and characterization techniques, and to discuss new applications of such techniques as well as their current limitations. More than 60 papers were submitted to this special issue, witnessing the importance of this area of research in today’s academic and industrial environment. With the help of more than 120 dedicated reviewers, we selected eight top-quality articles. A first subset deals with novel techniques for traffic classification and analyzes their accuracy. More precisely, ‘‘A non-linear, recurrence-based approach to traffic classification” presents a new, recurrent-plot based mechanisms for packet-level traffic classification; ‘‘Characterizing network traffic by means of the NetMine framework” describes a novel methodology to leverage advanced data-mining techniques to classify traffic at the packet level; ‘‘Efficient Application Identification and the Temporal and Spatial Stability of Classification Schem” analyzes the stability of statistical classification results considering different periods of time. A second set of articles deals with optimization mechanisms that apply to traffic classification, such as those obtained by considering new hardware architectures ‘‘Architecture and Scalability of a High-Speed Traffic Measurement Platform with a Highly Flexible Packet Classification” or efficient look-up algorithms for large databases of filtering rules ‘‘Scalable Packet Classi-