A Bio-inspired Method to Realize Fault-tolerance Online

• Published in: 2019 International Conference on High Performance Big Data and Intelligent Systems (HPBD&IS) pp. 169 - 175
• Main Authors: Xiao, Yi, Lu, Huaxiang, Chen, Gang, Mao, Wenyu
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.05.2019
• Subjects: bionic hardware; Circuit faults; distributed routing algorithm; Fault tolerance; Fault tolerant systems; Hardware; Heuristic algorithms; Immune system; intelligent system; Routing; self-diagnosis; self-healing
• ISBN: 9781728104669; 1728104661
• DOI: 10.1109/HPBDIS.2019.8735482

A multi-layered array inspired by biological processes and structures to support the self-diagnosis and self-healing of circuits is proposed. A dedicated distributed dynamic maximum flow (DDMF) algorithm is carefully designed to run on the array with an acceptable cost of time and hardware resources which makes self-adaptive online recovery possible. The array consists of three layers logically. Cells of the Programmable Logic Layer can be configured to different circuit functions. The Immune Layer, which cooperates with cells to detect and locate faults, is inspired by the human body's powerful immune system. Reconfiguration Manage Layer executes the dynamic routing algorithm to map the circuit function to spare cells automatically to realize online recovery. A fault-tolerance hardware platform is built based on the array. As the simulation and synthesis results show, the array performs well as a programmable platform and has the ability of distributed online self-diagnosis and self-healing.