Network Coding as a Performance Booster for Concurrent Multi-Path Transfer of Data in Multi-Hop Wireless Networks

• Published in: IEEE transactions on mobile computing Vol. 16; no. 4; pp. 1047 - 1058
• Main Authors: Arianpoo, Nasim, Aydin, Ilknur, Leung, Victor C. M.
• Format: Magazine Article
• Language: English
• Published: Los Alamitos IEEE 01.04.2017; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: algorithm/protocol design and analysis; Buffers; Coding; coding and information theory; Computer simulation; Data transfer; Data transfer (computers); Encoding; Machine learning; Markov processes; Network coding; Packets (communication); Protocol (computers); Protocols; Queues; Queuing theory; Receivers; Reliability; SCTP-CMT; Spread spectrum communication; State of the art; TCP-IP; Transport; wireless; Wireless communications; Wireless networks
• DOI: 10.1109/TMC.2016.2585106
• ISSN: 1536-1233

The emerging use of multi-homed wireless devices along with simultaneous multi-path data transfer offers tremendous potentials to improve the capacity of multi-hop wireless networks. The use of simultaneous data transfer over separate disjoint paths in multi-hop wireless networks to increase network capacity is a less explored subject, mainly because of the challenges it triggers for the reliable transport layer protocols. Reliable transport layer protocols generally use packet sequence number as a mean to ensure delivery. As such, the out-of-order packet arrival in reliable transport layer protocols triggers receiver buffer blocking that causes throughput degradation and prevents the reliable multi-path transport layer protocol to realize its vast potential. This paper focuses on integrating network coding with a reliable multi-path transport layer protocol to resolve the receiver buffer blocking problem. We propose an adaptive network coding mechanism to desensitize the receiver against packet reordering and consequently eliminate the receiver buffer blocking problem. Our state-of-the-art network coding scheme uses a combination of Q-learning and logistic regression for rare data events to control the number of redundant packets based on the network dynamics. We confirmed the veracity of our proposed scheme by a queuing theory based mathematical model. Moreover, the effectiveness of the proposed scheme is demonstrated through simulations and testbed experiments.