Methods of network traffic signs extractions based on signal processing researched

The next scientific seminar on "Methods of network traffic signs extractions based on signal processing" was held at the Institute of Information Technology of ANAS.

Senior Research Fellow, Ph.D. Ramiz Shikhaliyev presented the report, highlighted the relevance of the topic and pointed out that application of signals processing methods is one of the prospective trends in the classification of extracting network signs.

Today there are numerous peer-to-peer applications (P2P), social networks, streaming video services, online games, and so on, he said. This will increase the number of users and change their behavior. As a result, the volume of Internet traffic increases and the nature changes.  It is difficult to provide Productivity and security of networks, as well as applications and quality of service (Quality of Service). Effective monitoring is required to ensure that the network is normal and secure, and accurate identification of network traffic is important. Traditionally, network traffic identification is based on the analysis of network traffic signs. These features include certain attributes of packages: port numbers, IP address of sender and recipient, types of applications and protocols, contents of packages, statistical signs of traffic, etc. However, with the emergence of new network applications, the nature of the existing networking features varies, and the use of other features for network traffic identification appears.

"One of the most important research areas in network traffic identification is the classification of network traffic," the reporter said, the purpose of the classification is to set up a classification model for predicting unspecified network traffic based on a large number of trafficking tracks. R. Shikhaliyev said that the accuracy of the classification depends greatly on the scope and relevance of the classification attributes derived from network traffic. Development of methods for removing new classification signs from network traffic is one of the topical issues to ensure high accuracy of the classification.

The speaker provided detailed information on approaches to modeling the various network traffic characteristics as digital signals, and highlighted comparative analysis of the existing signal processing methods from network traffic to the issue of symptoms, analyzed their advantages and disadvantages.

At the end of the lecture, the views were exchanged, questions were answered. Head of department, PhD in technical sciences, associate professor Yadigar Imamverdiyev  recommended to continue research in this direction.

© All rights reserved. Citing to www.ict.az is necessary upon using news