Implementation of a hardware/software platform for real-time data-intensive applications in hazardous environments

In real-time data-intensive applications, the simultaneous achievement of the required performance and determinism is a difficult issue to address, mainly due to the time needed to perform I/O operations, which is more significant than the CPU processing time. Additional features need to be considered if these applications are intended to perform in hostile environments. In this paper, we address the implementation of a hardware/software platform designed to acquire, transfer, process and store massive amounts of information at sustained rates of several MBytes/sec, capable of supporting real-time applications with stringent throughput requirements under hazardous environmental conditions. A real-world system devoted to the inspection of nuclear power plants is presented as an illustrative example.