A piggybacking design framework for read-and download-efficient distributed storage codes

We present a new piggybacking framework for designing distributed storage codes that are efficient in the amount of data read and downloaded during node-repair. We illustrate the power of this framework by constructing explicit codes that attain the smallest amount of data to be read and downloaded for repair among all existing solutions for three important settings: (a) codes meeting the constraints of being maximum distance separable (MDS), high-rate, and having a small number of substripes, (b) binary MDS codes for all parameters where binary MDS codes exist, and (c) MDS codes with the smallest repair-locality. In addition, we show how to use this framework to enable efficient repair of parity nodes in existing codes that are constructed to address the repair of only the systematic nodes. The basic idea behind this framework is to take multiple stripes of existing codes and add carefully designed functions of the data of one stripe to other stripes. Typical savings in the amount of data read and downloaded during repair are 25% to 50% depending on the choice of the system parameters.