An LDPC Approach for Chunked Network Codes

作者：Tang, Bin^*; Yang, Shenghao

来源：IEEE/ACM Transactions on Networking, 2018, 26(1): 605-617.

DOI：10.1109/TNET.2017.2787726

• Random linear network coding
• chunked network codes
• LDPC
• belief-propagation Gaussian-elimination decoding
• degree-distribution optimization

Efficient communication through a multi-hop network with packet loss requires random linear network coding schemes with low computation cost and high throughput. In this paper, we propose a low-density parity-check (LDPC)-based framework for constructing chunked code, a variation of random linear network code with low encoding/decoding computational cost and small coefficient vector overhead. Two classes of chunked codes with LDPC structures, named uniform LDPC-chunked codes and overlapped LDPC-chunked (OLC) codes, are studied under a general chunk transfer matrix model. ULC codes achieve rates close to the optimum and perform better than existing chunked codes that employ parity-check constraints. OLC codes are overlapped chunked codes, where it is not necessary to generate new packets for encoding, and demonstrate much higher rates in certain scenarios than the state-of-the-art designs of overlapped chunked codes. These results justifies the feasibility of this LDPC approach for communication through multi-hop networks with packet loss.

• 出版日期2018-2
• 单位南京大学; 香港中文大学（深圳）