China
Africa
Explore chapters and articles related to this topic
Classification of P2P file-sharing traffic using heuristic based and statistical based techniques
Published in Rajesh Singh, Anita Gehlot, Intelligent Circuits and Systems, 2021
The oldest and the simplest network traffic classification is port-based classification, which classifies the traffic by analysing the TCP/UDP port number of the packet header. This is due to that fact that there are well-known port numbers ranging between 0 and 023 that are assigned by IANA [5] to various protocols such as HTTP, SMTP, DNS, FTP, etc. This technique is fast, easy to implement and hence is most often used to build certain rules in the firewalls and access control lists [6]. However, the port-based technique cannot be used to classify all kinds of traffic flows since there exists some protocols such as peer-to-peer (P2P) and FTP (operating in passive mode) which mostly use random or ephemeral port numbers for communication and hence they do not usually function on their default port numbers. In addition to this, many P2P applications masquerade the traffic to avoid their detection. The authors in [7] and [8] found that only 30–70% of the traffic could be detected by using port-based classification technique.
Smart Healthcare in Smart Cities
Published in Lavanya Sharma, Towards Smart World, 2020
In recent times, cloudlets developed elsewhere as data centers enhanced the latency and bandwidth of the network. They enhance the quality of service (QoS) of an interconnected healthcare network through a decrease of the latency and enhancement of the network capacity, connectivity, and fault tolerance. Traffic classification helps in the identification of the diverse procedures and applying it inside the network, enhancing safety by identification and neutralizing malicious packages. Cloudlets, as a specific part of the central cloud, are mobile edge-based self-managing technology, which collect and integrate data of Wi-Fi or mobile base stations. It consists of the assembly of devices, cloudlets, and clouds. Networking of 5G network stations with cloudlets will decrease the latency to value < 1 ms, because they can mutually communicate during recovering possible faults and respond to the various user demands. In processing data, the particular position has deep neural networks (DNN) based on deep learning (DL) supported by the help of a specific softer, which consists of multiple processing sheets that enable learning and recognition of various specific cases by using a back-propagation algorithm for indication of how changes of computer internal parameters can improve information in the specification sheet from the data of the previous sheet, discovering all complex connections in and between the layers [49, 50].
A Fast Traffic Classification Method Based On Sdn Network
Published in Amir Hussain, Mirjana Ivanovic, Electronics, Communications and Networks IV, 2015
Network traffic classification plays an important role in the modern network security and management (LIU 2008) especially in the application classification field. Classification results are widely used in network planning, service quality analysis, intrusion detection, user fees and other network management. On the other hand, it can be applied to the user's behavior analysis. For service providers, they can better understand user behaviors in order to provide more personalized service to enhance user satisfaction using flow classification techniques.
An Optimal Reinforced Deep Belief Network for Detection of Malicious Network Traffic
Published in IETE Journal of Research, 2023
Recently, the success of the Internet of Things (IoT) is seen in many fruitful applications. Through the steady growth of smart cities, many numbers of IoT systems are taken and there is a rapid growth in responsive data traffic [1]. Traffic classification is one of the significant functions in protecting against traffic attacks and confirming network security. However, the quality of service (QoS) and work efficiency is enhanced through various network traffic classifications [2]. With multimedia traffic and normal data, the smart applications in IoT generate the latest relationships and traffic kinds that contain bundles of active nodes such as actuators and sensors. The IoT arrival creates important challenges particularly for accessing the networks and networking in common [3]. For cooperating with numerous wireless devices collection, ultra-low-latency nonstop communication ability is provided by the 5G [4]. The data sharing and interconnection are provided by IoT between systems, buildings, and vehicles through different sensors along with the 5G growth and utilized in various services like smart buildings, e-commerce, and health care.
Hybrid feature learning framework for the classification of encrypted network traffic
Published in Connection Science, 2023
To have a better understanding of the ongoing network traffic, it is needed that we have a framework capable of capturing the network traces and providing a characterisation of it. Traffic classification (TC) is the medium of characterising the network data packets and labelling/segmenting them accordingly. A TC can be branched into two types: Classification based on Applications and Classification based on the file contents. Both these methods provide us with adequate knowledge of the network traces that we deal in our day-to-day life. In this paper, we have proposed a novel framework combined with a Deep Learning Neural Network that performs both types of classification mentioned above.