General Fractional Repetition Codes From Combinatorial Designs

Fractional repetition codes are a class of distributed storage codes dedicated to optimize the node repair performance, i.e., they enable uncoded exact repairs with minimum bandwidth consumption. Recent studies extend the application of such codes to heterogeneous storage networks, where the resultant codes are referred to as general fractional repetition (GFR) codes. In this paper, we propose some new constructions of GFR codes based on the combinatorial structures including packings, coverings, and pairwise balanced designs. Due to the mathematical properties of employed designs, the proposed constructions support a larger set of code parameters compared with existing construction methods.