Energy efficient data survivability for WSNs via Decentralized Erasure Codes

Designing reliability solutions for WSNs poses intricate challenges due to limitations in processing power and available energy. Such networks are often deployed in harsh and inaccessible environments and are therefore required to be highly reliable. However, reliability normally translates to redundancy in hardware and other resources implying both complexity and higher costs. In this study, we consider data survivability in WSNs. We present a data-centric framework based on Decentralized Erasure Codes (DEC) to increase the likelihood of data survivability in case of sensor nodes failure. The proposed framework enables network engineers to estimate the redundancy in hardware and data to achieve a given data survivability level. We also show two approaches to reduce the energy requirements of the proposed coding scheme using Random Linear Network Coding (RLNC). In addition to being decentralized, the proposed schemes are low in complexity requiring only binary coding over F₂. We evaluate the performance of the proposed schemes by simulations and compare them to schemes with no network coding.