APPLICATION OF THE SLOW START MECHANISM IN NOISY CHANNELS OF IEEE 802.11 WIRELESS COMPUTER NETWORKS
V. S. Khandetskyi, Genadiy POLUKHIN · Information Technology Computer Science Software Engineering and Cyber Security · 2025
Purpose of the work. The purpose of the work is to modify the slow start algorithm used in fiber-optic computer networks operated under the control of the TCP transport protocol in conditions of transmission without noise or in the presence of insignificant noise, to effectively apply the modified algorithm in wireless IEEE 802.11 networks in conditions of significant noise intensity. Methodology. To reduce the overhead of transmitting information in IEEE 802.11 wireless networks, frame transmission is used in blocks. In wireline networks operating under the control of the TCP protocol, a mechanism that combines multiplicative and additive increases in the size of the congestion window is used to avoid packet loss. After an exponential increase in the number of packets, at some point the slow start threshold will be exceeded. Additional packets that enter the network get into congestion, and packet loss occurs. Therefore, after the threshold is exceeded, the window is halved and TCP switches to additive window growth. The general idea is that the TCP connection should work for as long as possible with a window size close to optimal one – not too small, so that the throughput is not low, and not too large so that congestion does not occur. Scientific novelty. In the proposed article, we use a discrete-time Gaussian memoryless channel as a channel for transmitting frame blocks from the station to the access point of the infrastructure domain. In such a channel, bit errors are independent and equally distributed over the bits of information of the transmitted block. To ensure the effective operation of the slow start algorithm in wireless IEEE 802.11 networks under conditions of significant noise intensity, a principle is proposed for determining the threshold value depending on the number of a given size frames in the transmission blocks and the level of external noise, which is determined by the bit error rate in the blocks. The algorithm is provided with the property of adapting to the noise level, and the number of frames in the transmitted blocks is automatically regulated by the algorithm. Conclusions. A modified dynamic algorithm for operation of the infrastructure domain of a wireless network using probabilistic characteristics of the block transmission process has been formalized. The domain throughput has been estimated depending on the noise intensity and the lengths of frames in blocks. The quantitative characteristics of the throughput reduction with increasing noise intensity and decreasing frame length in blocks is calculated. The number of frames in blocks is automatically regulated by the algorithm depending on the noise intensity.