A Low Overhead Error Correction Algorithm Using Random Permutation For Sram Pufs

Static Random Access Memory-based Physically Unclonable Function (SRAM PUF) is frequently used in cryptographic applications such as key generation and IP protection because of its low cost, simple operation and high security features. The stability of PUF response is susceptible to environmental noise, so it requires the assistance of error correction algorithms when used as a key or ID. However, the actual error correction capability of the theoretically selected Error Correcting Codes (ECC) is always lower than expected. In this paper, we explore the specific reasons why SRAM PUF cannot use the theoretically selected ECC algorithm directly. In addition, an efficient and concise preprocessing method for random permutation is proposed to disturb the original position of unstable bits in the SRAM PUF response, thus confusing its instability distribution. Our experimental results show that the processed SRAM PUFs can recover the response sequence stably without increasing ECC's error correction capability, which effectively saves the resource consumption of error correction circuit.