Update and Repair Efficient Storage Codes with Availability via Finite Projective Planes

Update performance is a common concern in modern distributed storage systems. In this work, we construct explicit update-efficient codes via finite projective planes, also having efficient local repair with availability, and a short description. We compare to other existing solutions, including block codes from convolutional codes.We analyze the repair behavior of the codes via analogy with decoding of LDPC codes over an erasure channel, and the performance of the distributed storage system based on the codes involving updates and repairs.