relation: https://khub.utp.edu.my/scholars/8975/
title: Biologically inspired network on chip fault tolerant algorithm using time division multiplexing
creator: Sethi, M.A.J.
creator: Hussin, F.A.
creator: Hamid, N.H.
description: Biologically inspired solutions are a novel way of solving the complex and real world problems. Due to the advanced nanoscale manufacturing processes and the complex communication requirements of the processing elements (PEs) various faults have occurred on NoC. The complexity and communication requirement of the NoC has also increased due to the heterogeneous devices. To support the complexity of NoC, the physical device sizes are scaled down, which have contributed to faults. Various fault tolerant techniques have been proposed in the literature to address the temporary faults. But all these algorithms have drawbacks in terms of adaptiveness and robustness. Bio-inspired NoC using Time division multiplexing (TDM) is based on the characteristics of biological brain. The technique is fault tolerant as it detects and bypass the faulty interconnects. With the help of TDM, multiple connections are possible between multiple sources and multiple destinations, which efficiently utilize the NoC bandwidth between PEs. To the best of our knowledge, TDM based bio-inspired NoC is the first algorithm to address the fault tolerance using the TDM approach as the average packet latency is increased by 2.45, while the average bandwidth and throughput is reduced by 1.86 and 14.05 respectively during the recovery of faults. Also, the accepted traffic (flit/cycle/node) of the proposed bio-inspired technique is better than traditional fault tolerant techniques by 68.45. © 2016 IEEE.
publisher: Institute of Electrical and Electronics Engineers Inc.
date: 2017
type: Conference or Workshop Item
type: PeerReviewed
identifier:   Sethi, M.A.J. and Hussin, F.A. and Hamid, N.H.  (2017) Biologically inspired network on chip fault tolerant algorithm using time division multiplexing.  In: UNSPECIFIED.     
relation: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85011966498&doi=10.1109%2fICIAS.2016.7824058&partnerID=40&md5=0600bf30b9673b3432e7b78aed69759f
relation: 10.1109/ICIAS.2016.7824058
identifier: 10.1109/ICIAS.2016.7824058