Seminar: “FPGA-based fault injection”
Next Thursday 31/05/2018 at 12h30, researcher Jose Luis Nunes (University of Coimbra) will provide a talk describing his research on FPGA-based Fault Injection. Find here after the information concerning this seminar:
Title: FPGA-based Fault Injection
Summary: Reconfigurable embedded devices built on SRAM-based Field Programmable Gate Arrays (FPGA) are being increasingly used in critical embedded applications. Its susceptibility to Single Event Upsets (SEU) requires the use of fault tolerant designs, for which fault injection is one of the most accepted verification techniques.
In this talk implementation details of FIRED, a fault injector targeted at SRAM-based FPGAs (Virtex-5) for dependability evaluation of critical systems are presented. This tool is able to perform hardware fault injection in real-time, by inserting bitflips at the SRAM cells through Partial Dynamic Reconfiguration (PDR).
The architecture of an updated version of the tool, targetting current state-of-the-art devices (Xilinx 7-series) is also discussed. This new tool will take advantage of Soft Error Mitigation (SEM) core to support the fault injection.
CV of the speaker: José Luís Nunes is a professor at Coimbra Polytechnic where he teaches Operating Systems, Digital Systems and Computer Architectures. He is a member of the Software and Systems Engineering (SSE) research team at Center for Informatics and Systems of the University of Coimbra (CISUC), enrolled in the PhD program. His current research topics include dependability, fault injection, FPGAs and real-time embedded system.
Seminar at LAAS
This week, Juan Carlos Ruiz has been teaching the entitled seminar “Statistical Fault Injection: When is it enough in robustness assessment?”
“Simulation-based fault injection is commonly used to assess the robustness of hardware components modelled using Hardware Description Languages (HDL). The current complexity of modern circuits usually makes not feasible the consideration during experimentation of all possible combinations of fault models, targets, and times. By assuming a confidence interval and error margin, statistical fault injection exploits the principle of statistical sampling to reduce the number of experiments while keeping the results representative of the whole population of fault injections. Since the percentage of injected faults leading to failure is a priori unknown, such number of experiments is usually determined by selecting the value maximizing the sample size. This presentation argues that this conservative assumption leads to a worst-case scenario that can be improved. It proposes an new iterative approach to progressively adjust the number of experiments by estimating the percentage of those leading to failure and the error of such estimation. This proposal provides new means to decide when to stop a fault injection campaign and to estimate the error existing in the results finally reported.“