Research On The Techniques Of Data Communication Based On Network Coding

With the fast development and extensive application of the communication networktechnology, the demands for network data communication quality become higher and higher.Nowadays, it has become an important issue for network communication on how to improve theperformance of network data transport by more efficiently utilizing network resources under theexisting network conditions. Network coding, as information processing technology whichcombines coding with routing, provides a new solution to the issue. It allows networkintermediate nodes to perform the encoding operation during network data transport. This makesnetwork data transport no longer limited to the store and forward operation in traditional networkdata processing. Network coding can significantly improve the performance of network datatransport, such as improving network throughput and the utilization of network bandwidthresources, balancing network traffic-load and enhancing network robustness. Network coding isconsidered to be a major breakthrough in the field of information processing and transmission inthe21st century. And it has important theoretical value and broad application prospect.This dissertation attempts to make a research on how to improve the efficiency of networkdata transport with network coding. The research mainly focuses on such important issues and thecorresponding methods as the position of network coding in the existing network protocol stack,the optimization of network coding schemes and the minimalization of negative effects. That is,three aspects are studied, the combination of network coding with the existing network protocolsin each network layer, the maximization of network coding gain based on the control over the dataflow pattern and the delay control and reliability guarantee in network coding. And this makesnetwork coding achieve the maximal advantages in combining with the existing networkprotocols and minimizing the negative effects. The contributions of this dissertation are asfollows:First, the network cross-layer protocol structure with network coding is studied in detail. It isan important aspect to study the combination of network coding with the existing networkprotocols in each network layer. And this exploration will perfect the research on network codingand promote its application. Based on this analysis, the dissertation proposes a cross-layer protocol model based on network coding with joint design on the key parameters in each protocollayer. The linear programming function related to the network coding gain is constructed based onthe important parameters in each protocol layer. And the parameters are set with optimizationmanner by obtaining the optimal solution of the linear programming function. And this willachieve the optimal performance of the whole network system based on network coding. Thedistinct features are manifested by theoretic analysis.Second, the dissertation investigates the network coding model based on data rate matchbetween the coding data flows. In order to make full use of network coding, it is an important partto improve the network coding opportunity. And the network coding opportunity can be increasedby choosing the encoded nodes to construct the path in coding-aware routing protocols. In view ofthis, the dissertation proposes the optimization mechanism for network-coding based onrate-matching. The mechanism achieves data rate match of coding data flows in coding-awarepaths based on solving the linear programming function related to the data transport rate of dataflows. And this improves the coding opportunity to a large extent from the perspective of data ratematch and control. Meanwhile, the dynamic rate match adjustment is achieved by the informationfeedback process based on the statistical analysis on the decoding delay of data generationreflecting the data decoding state in destination nodes, which enhances the dynamic adaptability.In addition, the mechanism mitigates the data transport congestion in the process of enhancingnetwork coding opportunity and balancing network traffic-load. Theoretic analysis and simulationshow the validity and reliability of the model.Third, the research has been done on the dynamic network coding with delay constraint. Thetime for packets arriving at network intermediate nodes is different due to the different datatransport rates of data flows and the unsteady state of network links, which makes some packetswait for their coding matching packets. Meanwhile, in order to decode the packets in the currentdata generation, the destination nodes need to wait till receiving enough coding packets. Inaddition, network intermediate nodes frequently perform the encoding and decoding operations.All of these processes increase the data transport delay in network coding. Therefore, thedissertation proposes network coding model based on packet transport delay constraint withreal-time measurement on the packet per-hop accumulative transport delay. The modeldynamically determines the packet coding opportunity according to the accumulative packet transport delay. And it schedules the packets with priority in terms of the priority of data flows towhich the packets belong and the current accumulative transport delay of packets, whichguarantees the fast forwarding for packets with high service priority. Meanwhile, in order to speedup the transport of each data generation and avoid redundant data transmission, the modelintroduces the per-hop feedback transmission scheme on the acknowledgment packets. Thisachieves the effective control on data transport delay to the utmost extent in network coding andthe validity is tested in simulation experiments.Fourth, the coding packet forwarding mechanism based on the neighbor cooperation isproposed in the dissertation. In network coding, the coding packets display some independencerelatively. Therefore, in order to enhance the data transport reliability of the coding packets, it isan important issue to guarantee the coding packet transport in the paths with high communicationquality. The mechanism introduces the packet forwarding probability parameter and constructs theforwarding node set according to the packet forwarding probability. And based on the packetforwarding credit parameter of the nodes in the forwarding node set, the mechanism constructsseveral paths with high packet transport reliability. And in this way, the network data transportreliability is guaranteed by combining the data forwarding probability statistic with multi-pathrouting mechanism. The validity of the proposed mechanism is performed through simulation.