This paper presents a stochastic model for a Multi-operational-mode TCP variant optimised to effectively utilise the bandwidth in high-speedlong-delay networks. Particularly, a model for a TCP variant that increases its rate as power function of the current congestion window, and uses Multiplicative and subtractive decrease to reduce the congestion window upon packet loss. The power function is adapted dynamically using powers of {0. 5, 1, 2} according to the level of congestion in the network. The proposed model can be generalised to other powers; it captures the dynamics of the congestion window size evolution and sending rate under the influence of random packet loss by providing a closed-form expression for both the congestion window size and the normalised sending rate. We show that depending on the increase/decrease rules adopted for the congestion window size evolution, the congestion window size can be modelled either as function of exponential random variable or a Markov chain. Simulation results validate the analytical model and show that the Multiplicative decrease factor has more impact on the rate especially when a quadratic – instead of linear – increase rule is adopted under heavy packet loss rate. In addition to that, results show that the initial value of additive increase when a square-root rule is adopted has minimal effect on the average congestion window size.

Standard TCP is the de facto reliable transfer protocol for the Internet. It is designed to establish a reliable connection using only a single network interface. However, standard TCP with single interfacing performs poorly due to intermittent node connectivity. This requires the re-establishment of connections as the IP addresses change. Multi-path TCP (MPTCP) has emerged to utilize Multiple network interfaces in order to deliver higher throughput. Resilience to link failures can be better supported in MPTCP as the segments’ communication are maintained via alternative interfaces. In this paper, the resilience of MPTCP to link failures against several challenges is evaluated. Several link failure scenarios are applied to examine all aspects of MPTCP including congestion algorithms, path management, and subflow scheduling. In each scenario, the behavior of MPTCP is studied by observing and analyzing the throughput and delay. The evaluation of the results indicates MPTCP resilience to a low number of failed links. However, as the number of failed links increases, MPTCP can only recover full throughput if the link failure occurs on the server side. In addition, in the presence of link failures, the lowestRTT MPTCP scheduler yields the shortest delivery time while providing the minimum application jitter.

As wireless networks with high data rate get widely deployed, improving the performance of TCP over these networks plays vital role. Wireless link losses have dramatic adverse impact on TCP performance due to the difficulty in distinguishing the congestion losses from wireless link losses. We first, studied the various existing Bandwidth Estimation Algorithms over the wireless networks. In this paper, we propose an enhanced bandwidth estimation technique with congestion control algorithms to improve TCP performance over wireless link with random loss. We propose, a new technique called TCP- BWIW to estimate the bandwidth and the congestion window is set accordingly. The performance of TCP-BWIW is analyzed, and it mitigates the Network congestion.

Early detection of skin lesions is essential for the success of treatment depending on the earliest possible detection of skin cancer lesions. Segmentation of skin cancer lesions is one of the most important early steps. In this regard, classic U-Net which is based on deep neural networks is the most popular architecture for medical image segmentation. However, the classic U-Net architecture lacks certain aspects. In this approach, we proposed a lightweight model designed to minimize memory usage in the deeper network layers and to reduce training and testing time. We achieved this by leveraging Multi-Level Blocks, which exclusively utilized 3x3 convolution operations. Additionally, we have utilized Multiple convolutions to facilitate the transfer of information from the encoding to the decoding stage. This approach aims to minimize the semantic gap between the two stages. We have termed this information transfer path the encoder-decoder path. Our method has demonstrated outstanding performance in key metrics when tested on the PH2 dataset and has shown superior performance in terms of Accuracy and Jaccard Index on the ISIC-2017 dataset compared to the latest methods reported in existing publications. The Multi-path U-Net method effectively recognizes and precisely segments complex features such as weak boundaries, shape, and color irregularities, and Multi-part lesions with diverse color intensities.