Improving Reliable Transmission Throughput with Systematic Random Code

Rateless erasure code (REC) is an erasure code, where the encoder generates a potentially infinite number of encoded symbols and the original message can be reconstructed from a sufficient number of correctly received packets. Many REC-based transmission protocols have been proposed for improving network throughput in lossy channel. However, state-of-the-art RECs (such as LT code and Raptor code) are not efficient for transmitting short messages. Recent studies suggest that network traffic is characterised by bursts of short messages and thus existing transmission protocols do not benefit from the gains of deploying REC. In this paper, we propose an REC-based transmission protocol, namely UDP-RC, which integrates the simplicity of UDP and strength of systematic Random code suited to network traffic with short messages. It attains high throughput by transmitting short messages reliably with lower overheads over lossy channel. We experimentally show that UDP-RC achieves at least 50% higher throughput and maintains more stable throughput compared to TCP (Transmission Control Protocol) and UDT (UDP Data transfer) protocol under both ideal and lossy channel conditions.

[1]  Vasile Bota,et al.  Analysis of a transport protocol based on rateless erasure correcting codes , 2010, Proceedings of the 2010 IEEE 6th International Conference on Intelligent Computer Communication and Processing.

[2]  Bryan Ng,et al.  Developing a traffic classification platform for enterprise networks with SDN: Experiences & lessons learned , 2015, 2015 IFIP Networking Conference (IFIP Networking).

[3]  Wei Ding,et al.  Comparative Research on Internet Flows Characteristics , 2012, 2012 Third International Conference on Networking and Distributed Computing.

[4]  Bryan Ng,et al.  Cluster-centric medium access control for WSNs in structural health monitoring , 2015, 2015 13th International Symposium on Modeling and Optimization in Mobile, Ad Hoc, and Wireless Networks (WiOpt).

[5]  Alex C. Snoeren,et al.  Decongestion Control , 2006, HotNets.

[6]  Robert L. Grossman,et al.  UDT: UDP-based data transfer for high-speed wide area networks , 2007, Comput. Networks.

[7]  Steve Hranilovic,et al.  Short-Length Raptor Codes for Mobile Free-Space Optical Channels , 2009, 2009 IEEE International Conference on Communications.

[8]  Antonio Fernández,et al.  A game theoretic comparison of TCP and digital fountain based protocols , 2007, Comput. Networks.

[9]  David A. Maltz,et al.  Network traffic characteristics of data centers in the wild , 2010, IMC '10.

[10]  Koushik Kar,et al.  Loss-Tolerant TCP (LT-TCP): Implementation and experimental evaluation , 2012, MILCOM 2012 - 2012 IEEE Military Communications Conference.

[11]  Balázs Sonkoly,et al.  How to transfer flows efficiently via the Internet? , 2014, 2014 International Conference on Computing, Networking and Communications (ICNC).

[12]  Tomoaki Tsugawa,et al.  TCP-AFEC: An adaptive FEC code control for end-to-end bandwidth guarantee , 2007, Packet Video 2007.

[13]  Hiroyuki Ohsaki,et al.  Systematic rateless erasure code for short messages transmission , 2015, Comput. Electr. Eng..

[14]  Balázs Sonkoly,et al.  Data transfer paradigms for future networks: Fountain coding or congestion control? , 2013, 2013 IFIP Networking Conference.

[15]  Anghel Botos,et al.  FECTCP for high packet error rate wireless channels , 2010, 2010 8th International Conference on Communications.

[16]  Alexandre Proutière,et al.  Is the ''Law of the Jungle'' Sustainable for the Internet? , 2009, IEEE INFOCOM 2009.

[17]  John K. Zao,et al.  Design of optimal short-length LT codes using evolution strategies , 2012, 2012 IEEE Congress on Evolutionary Computation.

[18]  Mohit P. Tahiliani,et al.  TCP Kay: An end-to-end improvement to TCP performance in lossy wireless networks using ACK-DIV technique & FEC , 2015, 2015 IEEE International Conference on Electronics, Computing and Communication Technologies (CONECCT).