A model for the evaluation of fault tolerance in the FERMI system

Abstract

Experiments of high energy physics planned at the Large Hadron Collider (LHC) at CERN (CH) require digital data acquisition systems with high throughput. Such systems need also be fault tolerant to the permanent and transient faults induced by radiation, since they must be installed close to the experiment area. A model for the evaluation of the fault tolerance performances of the most significant VLSI devices designed for the construction of the FERMI system, a data acquisition system for the LHC, is presented. The model allows one to quantify the probability of the FERMI system to survive fatal failures, to increase mission time and to provide valid data for different implementations of fault tolerance.