China
Africa
Explore chapters and articles related to this topic
Mobile Satellite Channel Characterization
Published in Athanasios G. Kanatas, Athanasios D. Panagopoulos, Radio Wave Propagation and Channel Modeling for Earth–Space Systems, 2017
To understand the practical technical difficulties and limitations of MSS, a detailed knowledge and characterization of the underlying radio channel, under various propagation conditions and scenarios, is crucial. Then, efficient and reliable MSS can be designed and accurately tested, before their implementation. Using this knowledge, MSS can be designed to obtain optimal or near- optimal performance. This philosophy has been the driving force behind the research activity on all types of wireless communications systems. Although measurement campaigns are expensive, time-consuming, and difficult to carry out, conducting measurements and collecting measured channel data is a precondition for the successful validation of the results of preliminary theoretical efforts. Hence, the characterization of the satellite radio channel through real-world channel measurements has received the attention of many researchers.
Identification of Network Application Behaviors Hiding in HTTP Tunnels
Published in IETE Technical Review, 2021
Huiwen Bai, Guangjie Liu, Weiwei Liu, Jiangtao Zhai, Luhui Yang, Yuewei Dai
The number of packets in the HTTP flow is not fixed. It is determined by factors such as the size of the current transmission object, MTU, and MSS. Flow fragmentation is to divide each flow into several equal-length fragments, and each fragment is as a sample. Assuming that each HTTP tunneling flow contains only one type of behavior and the behavior distribution of each flow is uniform, namely, any continuous number of packets in the flow can reflect some characteristics of these behaviors. Flow fragmenting, on the one hand, is to make the packet amount of each sample consistent. On the other hand, it can reduce the computation overhead and improve the classification efficiency so as to achieve real-time identification requirements. The specific fragmenting rules are described as follows: Assuming there are n packets in a flow excepting handshake packets, the flow can be denoted as . Each flow is divided according to the fragment length :If , the flow is directly split to a fragment;If , the flow is split to fragments. They can be denoted as where . The size of the fragment will determine the amount of information in the fragmented sample. Too small fragments (such as 1 packet for 1 fragment) can affect the accuracy of classification. Excessive fragmentation can increase the computational cost of the classifier, and it may also make classification fuzzy. To obtain the most suitable model, we will conduct a comparison experiment under different fragment sizes in Section 6.2.