Research On Concurrent Multipath Transfer Based On Network Coding In Wireless Networks

With the development of society, people’s demand of high-quality and efficient wireless data transmission service becomes higher and higher. The extremely rapid development of wireless communication technologies and the improvement of mobile device intelligence have pushed the improvement of Concurrent Multipath Transfer (CMT). CMT has benefits of bandwidth aggregation, fault tolerance and load balance, which make it as one of the best way to enhance the performance of data transmission service. However, due to the path diversity in dynamic wireless environment, packets reordering exits seriously. The performance of data transmission is limited by the head-of-line blocking when the mobile device’s buffer is small. On the other hand, packets losses and retransmissions happen frequently because of the wireless paths’ unreliability. In some cases, retransmissions can’t remedy the disadvantages of packets losses especially for time-sensitive applications. Most of the existent solutions that solve the problem of packets reordering and losses are just depending on optimizing the data scheduling algorithm and retransmission scheme, which can’t solve these problems effectively and beforehand.According to the above problems, this essay proposes a novel Concurrent Multipath Transfer scheme based on Rolling Generation Network Coding (CMT-RG) to solve the problem of packets reordering and losses in heterogeneous wireless networks, so that it can provide a more efficient data transmission service. CMT-RG designs a data management module to process data according to handle the data diversity. Taking the dynamic wireless environment into consideration, a path quality model based on multi-parameters is proposed to measure and estimate the paths’quality. The key points of this essay is a Rolling Generation-based Network Coding algorithm (RGNC), which can dynamically adapt its coding scheme to the transmission status, aiming to solve the problem of packets reordering and losses. Simulation results show how CMT-RG outperforms exiting solutions in terms of transport throughput, retransmission and delay.