Efficient Erasure Code for Wireless Sensor Networks

Reliable communication in Wireless Sensor Networks is hard to achieve efficiently using methods in conventional systems like Internet. End-to-end retransmission is inefficient, or in some cases impossible. Erasure code is code with which we can reconstruct m original messages by receiving any m out of n code words (n > m). Erasure code provides efficient option for achieving reliability. Erasure code based on Reed-Solomon code is implemented. And further modifications were made to optimize for resource-constrained Wireless Sensor Networks mote. Systematic code is used to increase speed of decoding. Multiple independent code words are packed into a single packet to reduce the size of code word, and to reduce memory space and computation time eventually. Operation lookup table is used for quicker decoding. Memory usage and computing time depend on the size of code word. For most of practical usage in Wireless Sensor Networks, 4 or 8 bit long code word will be sufficient. Then moderate amount of memory is used, and computation is quick with operation table. Redundant information and flexibility in loss distribution give benefit to convergence routing, divergence routing (multicast), and point-to-point routing.