Title
Towards Zero Retransmission Overhead: A Symbol Level Network Coding Approach to Retransmission
Abstract
We present SYNC, a physical layer transmission scheme that drastically reduces the cost of retransmission by introducing network coding concepts to symbol level operation. It piggybacks a new packet on each retransmitted packet, and exploits the previously received packet (possibly with error) at the receiver to recover both the retransmitted packet and the piggybacked packet. The piggybacking is achieved through higher modulation, but it does not decrease the decodability of the mixed packets owing to the previously received packet at the receiver, which can be analytically shown. SYNC works independently of other PHY level performance boosting schemes such as channel coding and spatial diversity. The proof-of-concept SYNC prototype has been implemented on a software defined radio (SDR) platform. The measurement data shows that under the same channel condition SYNC achieves 110 percent and 42 percent median throughput gain over traditional retransmission and SOFT, respectively. We also show that SYNC can be used proactively where the feedback as to the success of the previous transmission is not available, such as in broadcast.
Year
DOI
Venue
2011
10.1109/TMC.2010.248
IEEE Trans. Mob. Comput.
Keywords
Field
DocType
spatial diversity,piggybacked packet,mixed packet,software defined radio,new packet,receiver,channel coding,zero retransmission overhead,percent median throughput gain,proof-of-concept sync prototype,temporal diversity,retransmission,symbol-level network coding,channel condition sync,receivers,towards zero retransmission overhead,throughput.,software radio,prototype,physical layer transmission scheme,phy level performance,symbol level network coding,retransmitted packet,network coding,decoding,throughput,synchronization,encoding,modulation
Linear network coding,Synchronization,Computer science,Retransmission,Network packet,Computer network,Real-time computing,Packet switching,Throughput,PHY,sync
Journal
Volume
Issue
ISSN
10
8
1536-1233
Citations 
PageRank 
References 
8
0.57
13
Authors
3
Name
Order
Citations
PageRank
Sangki Yun120516.77
Hyogon Kim249458.48
Kun Tan3135098.64