Generalized Piggybacking Codes for Distributed Storage Systems
Shuai Yuan, Qin Huang · 2016
This paper generalizes the piggybacking constructions for distributed storage systems by considering various protected instances and piggybacked instances. Analysis demonstrates that the proportion of protected instances determines the average repair bandwidth for a systematic node. The generalized codes cost much less repair bandwidth than the previous piggybacking codes by optimizing the proportion of protected instances. Asymptotically, the average repair bandwidth as a fraction of the original messages approaches zero instead of 50% as the number of parity check nodes tends to infinity.