Dependability Analysis of Fault Tolerant Systems Based on Partial Dynamic Reconfiguration Implemented into FPGA

Kastil, Jan; Straka, Martin; Mičulka, Lukáš; Kotásek, Zdenĕk

doi:10.1109/dsd.2012.40

Cited by 12 publications

(3 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The cumulative reward of this single Markov Reward Model (MRM) [17] is then used to evaluate the corresponding design option in terms of reliability, availability, safety and performability-area trade-off. Current work in this area [18,19,20,10] either separates the dependability analysis from performance/area, coverage analysis, or does not analyze such safety-critical appli-cations at an early design stage. Commercial tools for dependability analysis, such as isograph [21], concentrate mainly on the reliability and availability analysis and lack the support for performability evaluation.…”

Section: Contributions and Limitations Of Previously Reported Approachesmentioning

confidence: 99%

Formal analysis of SEU mitigation for early dependability and performability analysis of FPGA-based space applications

Hoque

Mohamed

Savaria

2017

Journal of Applied Logic

View full text Add to dashboard Cite

SRAM-based FPGAs are increasingly popular in the aerospace industry due to their field programmability and low cost. However, they suffer from cosmic radiation induced Single Event Upsets (SEUs). In safety-critical applications, the dependability of the design is a prime concern since failures may have catastrophic consequences. An early analysis of the relationship between dependability metrics, performability-area trade-off, and different mitigation techniques for such applications can reduce the design effort while increasing the design confidence. This paper introduces a novel methodology based on probabilistic model checking, for the analysis of the reliability, availability, safety and performance-area tradeoffs of safety-critical systems for early design decisions. Starting from the high-level description of a system, a Markov reward model is constructed from the Control Data Flow Graph (CDFG) and a component characterization library targeting FPGAs. The proposed model and exhaustive analysis capture all the failure states (based on the fault detection coverage) and repairs possible in the system. We present quantitative results based on an FIR filter circuit to illustrate the applicability of the proposed approach and to demonstrate that a wide range of useful dependability and performability properties can be analyzed using the proposed methodology. The modeling results show the relationship between different mitigation techniques and fault detection coverage, exposing their direct impact on the design for early decisions.

show abstract

Section: Contributions and Limitations Of Previously Reported Approachesmentioning

confidence: 99%

Formal analysis of SEU mitigation for early dependability and performability analysis of FPGA-based space applications

Hoque

Mohamed

Savaria

2017

Journal of Applied Logic

View full text Add to dashboard Cite

show abstract

“…The methodology takes particular radiation environments into account and is based on fault injection experiments. Kastil et al present a dependability analysis of their fault tolerant systems in [8]. Applications hosted on SRAM-based FPGAs are partitioned into functional units to which different redundancy configurations can be applied.…”

Section: Availability Analysis Methodologymentioning

confidence: 99%

Availability analysis for satellite data processing systems based on SRAM FPGAs

Siegle

Vladimirova

Ilstad

et al. 2016

IEEE Trans. Aerosp. Electron. Syst.

View full text Add to dashboard Cite

Abstract-This paper presents a novel methodology, which allows a systematic availability analysis of satellite payload data processing systems implemented on SRAM-based FPGAs. The methodology allows (i) comparison of different Fault Detection, Isolation and Recovery (FDIR) schemes and (ii) prediction of the expected system availability in a particular radiation environment. Furthermore, it advances the state of the art by analysing embedded Block RAMs and employing a novel fault injection algorithm that enables more complex stochastic models. The applicability of the method is demonstrated by a case study representing a high availability payload data processing application. Since Block RAMs are also taken into account, the availability prediction precision is greatly increased. It is shown that the reliability prediction for two border cases, in which the Block RAMs are either ignored or assumed to be fully susceptible, can differ as much as 75%. With the proposed Block RAM profiling tool it is possible to determine a realistic reliability figure that is eventually required for the accurate estimation of the system availability.

show abstract

“…Ostler et al present a reliability analysis of SRAM-based FPGAs in [4]. Kastil et al present a dependability analysis of their fault tolerant systems in [5]. Martin et al also use Markov chains in [6] to model the availability that can be achieved with different scrubber implementations.…”

Section: Introductionmentioning

confidence: 99%

Validation of FDIR Strategy for Spaceborne SRAM-Based FPGAs Using Proton Radiation Testing

Siegle

Vladimirova

Poivey

et al. 2015

2015 15th European Conference on Radiation and Its Effects on Components and Systems (RADECS)

View full text Add to dashboard Cite

Abstract-This paper presents a novel stream processor architecture for SRAM-based FPGAs that is specifically targeted at payload data processing and which employs innovative Fault Detection, Isolation and Recovery (FDIR) mechanisms to cope with failures caused by radiation effects. As part of this FDIR strategy, an availability analysis method is developed that is able to predict the steady state availability of a stream processor in a particular radiation environment. By means of an accelerated proton irradiation test campaign, both the FDIR framework and the availability analysis method are validated. First, it is demonstrated that the FDIR hardware and software components are capable to detect and recover from failures in a real radiation environment. Secondly, it is proven that the availability prediction provides accurate results. The real Mean Time Between Failures (MTBF) value measured during the beam test differs from the prediction by not more than 15.4% while the steady state availability by only 0.9%.

show abstract

Dependability Analysis of Fault Tolerant Systems Based on Partial Dynamic Reconfiguration Implemented into FPGA

Cited by 12 publications

References 23 publications

Formal analysis of SEU mitigation for early dependability and performability analysis of FPGA-based space applications

Formal analysis of SEU mitigation for early dependability and performability analysis of FPGA-based space applications

Availability analysis for satellite data processing systems based on SRAM FPGAs

Validation of FDIR Strategy for Spaceborne SRAM-Based FPGAs Using Proton Radiation Testing

Contact Info

Product

Resources

About