New rateless codes for receiver with limited memory

In this paper, we propose a new class of rateless codes which have improved intermediate performance using batched zigzag decoding over symbol erasure channels. When generating a batch, an encoder shifts selected input symbols before performing bit-wise XOR operation in order to make a batch zigzag decodable. Solving a batch by zigzag decoding algorithm gives improved intermediate performance. We show that the proposed codes outperform LT codes and zigzag decod-able rateless codes with respect to coding overhead when symbol erasure rate is low and available buffer is limited.