Reliability-aware placement and fault tolerant reconfiguration in FPGAs

With various applications of reconfigurable devices in harsh or hostile remote environments, reliability becomes an important factor to evaluate system performance. In this paper, an approach is presented to improve the reliability of FPGA-based system through optimal spare placement and fault tolerant (FT) reconfiguration. In FPGA placement stage, high reliability of spare distribution is equivalent to optimal redundancy allocation problem. A modified Simulated-Annealing (SA) algorithm which incorporates reliability to optimization objective is applied to solve the allocation problem. Based on graph theory, a reliability evaluation method suitable for complex system is also given. When a fault is detected, a fault tolerant reconfiguration which selects the swapped spare with the least affection of reliability is introduced to recover the fault. Simulation results show that this approach improved reliability and achieved less average moves in reconfiguration.