A Study On Fountain Code Based Data Exchange Strategy In Wireless Sensor Networks

Wireless Sensor Networks (WSN) is a class of wireless network that consists of numbers of discrete sensor nodes. Its structure is a typical distributed network system. As a data-centric distributed ad-hoc network, data gathering and storage are playing a important role in wireless sensor networks. Since that the deployment of wireless sensor networks is very convenient and powerful, it is often deployed in harsh environments for data gathering. Adding that resources such as CPU, memory and flash storage are limited in sensor nodes, conditions for WSN are also limited. So the guarantee of reliability and effectiveness of sensed data is a very crucial in wireless sensor networks. Also, these is a worthy study topic in the industry.Recently, a new fountain code based network coding technology is widely applied for data gathering and storage in wireless sensor networks. The difference between these and the traditional technology is that it applies coding to sensors’data before forwarding it to the next hop. Thus, when the receiver receives the encoded data from the sender, it can decode the encoded data with the agreed coding and decoding algorithm between sender and receiver. However, because the data sent to a receiver is encoded before sending out, a receiver has to receive enough encoded data before decoding all of the original data. Thus there may encounter some problems before a receiver decoding all of the original data:1. Steep effect of data decoding. Since a receiver can only receive a small little encoded packet firstly, thus they can’t be decoded successfully, which wastes a lot of time at the beginning. During the later period, a receiver can receive a large number of packets, which wastes a lot of bandwidth and may cause network congestion.2. To archive the best performance, network coding scheme assures that degree of packets according to a uniformly degree distributed function. Unfortunately, most of the existing coding scheme can’t archive this requirementAccording to these problems, this paper proposes a data exchange strategy in fountain code. Our contributions in this paper contain:1. Propose a new packet-centric coding algorithm, which makes nodes’data distribute in the network uniformly. With the simulated experiment result, we analyze the performance of this algorithm.2. For the proposed packet-centric coding algorithm, we propose a new random walk algorithm to make sure that the algorithm more efficient. Also, we simulate the algorithm and give an analysis of the performance of this algorithm. 3. Propose a new priority based data oriented exchanging strategy for the proposed packet-centric coding algorithm. Simulation result shows that it is more efficient when the collector starting collecting data along the edge of the network.