Analysis of Inter-flow Network Coding in Lossy Wireless Networks | ||
| The Modares Journal of Electrical Engineering | ||
| Article 9, Volume 16, Issue 2, 2016, Pages 62-71 PDF (819.48 K) | ||
| Authors | ||
| Alireza Shafieinejad* 1; Faramarz Hendessi2 | ||
| 1Department of Electrical and Computer Engineering, Tarbiat-Modares University, Tehran, Iran | ||
| 2Department of Electrical and Computer Engineering, Isfahan University of Technology, Isfahan, Iran | ||
| Abstract | ||
| This paper addresses the problem of inter-flow network coding for unicast sessions in lossy channel wireless networks. In spite of decreasing the number of transmissions, network coding intuitively increases the sensitivity of nodes to lost packets. First of all, coded packets carry more information than native packets and thus losing a coded packet prohibits a series of dependent nodes from decoding their intended packets. Secondly, for the scheme with opportunistic listening, it is necessary for some of the nodes to overhear the transmission of their neighbors. Thus, successful decoding requires overhearing of the corresponding packet(s) in addition to correct reception of unicast and broadcast transmissions. In this paper, we study the effect of lossy channel on the aggregate network throughput in the presence of network coding. We provided a linear programming formulation to compute the throughput performance of network coding for a general lossy wireless network. Further, we consider a retransmission mechanism for both unicast and broadcast. Our LP system supports both COPE and Star coding schemes. The advantages of the proposed NC schemes over the non-NC ones are shown through simulations and theoretical analysis. Results show that network coding can boost the capacity of wireless network up to 40% under lossy channel condition. | ||
| Keywords | ||
| Wireless network coding; Lossy wireless networks; Reliable broadcasting | ||
|
Statistics Article View: 75 PDF Download: 74 |
||
| Number of Journals | 45 |
| Number of Issues | 2,171 |
| Number of Articles | 24,672 |
| Article View | 24,384,175 |
| PDF Download | 17,531,371 |