Abstract:Locally repairable codes can effectively address the challenge of multi-node failure recovery in distributed storage systems. To overcome the limitations of existing locally repairable codes, such as inflexible fault tolerance and low code rates, a method for constructing locally repairable codes based on cyclic permutation matrices is proposed. This method utilizes cyclic permutation matrices to construct a parity-check matrix, from which the locally repairable codes are generated. Experimental validation shows that, compared to existing locally repairable codes, the constructed codes achieve optimal minimum distance and optimal code length, while offering higher code rates and more flexible availability parameter choices. However, their fault tolerance capability is constrained by the availability parameters. Furthermore, based on the aforementioned construction of locally repairable codes, another class of locally repairable codes is proposed by applying the Kronecker column-wise product operation to the parity-check matrix. This class of locally repairable codes achieves optimal minimum distance and enhances the flexibility of system fault tolerance. By adjusting the value of the local fault tolerance parameter δ , flexible fault tolerance capabilities can be achieved, meeting the application requirements of distributed storage systems for adjustable fault tolerance.