Adaptive Decoding for Fulcrum Codes
Vu Nguyen, Juan A. Cabrera, Giang T.K. Nguyen, Frank Gabriel, Christopher Lehmann, Stanislav Mudriievskyi, Frank H. P. Fitzek · 2018
Fulcrum codes, a variant of random linear network coding (RLNC), especially designed for heterogeneous environment. A Fulcrum encoder combines two codes operating on both small, such as binary and big, such as 28Galois fields. That allows the Fulcrum decoder to operate on either the smaller or greater field size, depending on whether it has a moderate or stronger computing power. However, setting on one particular Galois field is inefficient for end devices, especially in broadcasting scenarios when feedback to senders is normally impractical. We propose an adaptive decoding process that alters among Galois fields to decode. Our comprehensive evaluation shows that the proposed adaptive decoding significantly reduces the computation complexity at end devices while simultaneously maintaining a high decoding probability.