Abstract
Network coding has been investigated to improve the data transmission performance of wireless sensor networks (WSNs), and reduce data redundancy in data centers. In recent years, the amount of sensor data transmitted to data centers for analysis has increased dramatically, and traditional transmission schemes have encountered problems due to low transmission efficiency under conditions with massive data transmission. These problems are caused by the use of multiple codecs (including several coding and recoding codecs at the sensor nodes, one for decoding at the data gateway, one for coding for redundant storage and one for decoding to obtain original sensor data) and inefficient decoding processes. To solve these problems, we present an integrated fast data transmission scheme based on network coding. We consider the data transmission of WSNs and storage at data centers simultaneously so that codecs at data gateways can be avoided, and the encoded data can be stored directly in data centers. Furthermore, we design a gradient constraint flooding protocol based on network coding for data-intensive WSNs, and implement a fast, parallel decoding scheme for the data center. Our experimental results show that our scheme has the better performance in terms of rapid data delivery and decoding speed than traditional schemes.
Highlights
Since the start of the big data era, a large amount of data have been collected in wireless sensor networks (WSNs) [1] and transmitted to data centers for analyses
To further accelerate data delivery, we investigated a routing protocol based on network coding for data-intensive WSNs and parallel decoding schemes based on network coding
We divide the experiment into three parts: performance simulation of delivery time in gradient constraint flooding protocol (GCF), of decoding in fast parallel decoding scheme (FPD), and of the novel network architecture proposed in this paper
Summary
Since the start of the big data era, a large amount of data have been collected in wireless sensor networks (WSNs) [1] and transmitted to data centers for analyses. B. MAIN CONCEPT OF THE PROPOSED METHOD To solve the problem of multiple unnecessary encoding and decoding operations during transmission and storage, the key is to combine the data transmission in the network with the redundant storage at the data center. The coding for redundant storage is completed in the WSN, and sensor nodes share computing tasks with a data center, without increasing the number of calculations required. One may think that using data flooding would be the most effective solution as this strategy can ensure the full mixing of data Such flooding methods cannot control the number of packets sent across WSNs, the final coded packets received by the sink node, which results in a huge waste of bandwidth.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
