Argument Flow, a combination of classical Control Flow and dynamic Data Flow, is a very efficient program organization for fault -tolerant multiprocessing. Using this working principle, several fault-tolerance algoritms realizing a different degree of fault-tolerance are developed. Integrating all these methods in one global algorithm results in an optimum utilization of the hardware providing no more redundancy than required by the different processes. Both link and node failures are tolerated. It is even possible to recover from multiple simultaneously occurring errors.
© 2001-2025 Fundación Dialnet · Todos los derechos reservados