The Community for Technology Leaders
2015 International Conference on Big Data and Smart Computing (BigComp) (2015)
Jeju, South Korea
Feb. 9, 2015 to Feb. 11, 2015
ISBN: 978-1-4799-7303-3
pp: 150-156
Kyu-haeng Lee , Dept. of CSE, Seoul National University, 151-744, Seoul, Korea
Chong-kwon Kim , Dept. of CSE, Seoul National University, 151-744, Seoul, Korea
Soo-Hyung Lee , Embedded SW Research Department, Electronics Telecommunications Research Institute, 161 Gajeong-dong, Yuseong-gu, Daejeon, 305-700, Korea
Won-tae Kim , Embedded SW Research Department, Electronics Telecommunications Research Institute, 161 Gajeong-dong, Yuseong-gu, Daejeon, 305-700, Korea
ABSTRACT
The Data Distribution Service (DDS) based on the Publish and Subscribe (PUB/SUB) communication model provides a high flexible data transmission for real-time applications, yet it faces a significant challenge for achieving reliable multicast. The ARQ-based error control adopted in the DDS is effective only when the channel condition is moderately good, and is also vulnerable to feedback implosion. This paper describes a new reliable multicast scheme for DDS, called RCRM. First, RCRM employs a rateless code, so that it can obtain higher reliable multicast efficiency than the legacy ARQ-based DDS. Second, the novel feedback scheme of RCRM considerably enhances the rateless coding performance. In particular, the publisher finds appropriate feedback points to minimize the coding overhead. We implement RCRM on the MATLAB simulator, and conduct extensive simulations. The results show that RCRM outperforms other protocols in terms of the coding overhead and decoding completion time.
INDEX TERMS
Encoding, Decoding, Reliability, Protocols, Data communication, Optimization, Automatic repeat request
CITATION

K. Lee, C. Kim, S. Lee and W. Kim, "Rateless code based reliable multicast for data distribution service," 2015 International Conference on Big Data and Smart Computing (BigComp)(BIGCOMP), Jeju, South Korea, 2015, pp. 150-156.
doi:10.1109/35021BIGCOMP.2015.7072825
98 ms
(Ver 3.3 (11022016))