Designing the dependable multiprocessor space experiment

Abstract

During development of the Dependable Multiprocessor experiment, high-fidelity environment analyses and radiation tests were performed in order to demonstrate that a meaningful on-orbit experiment was probable. I. INTROD UCTION Environmentally Adaptiv e Fault Tolerant Computing (EAFTC) technology has been developed as part of NASA 's New Millennium Program (NMP) Space Technology S (STS) project [1-5]. The objective is to combine high performance COTS-b ased cluster processing with replication services and fault tolerant middleware in an architecture and software framework capable of support ing a wide variety of mission applications. The NMP phase of EAFTC development, known as the Dependable Multipro cessor (DM), is baselined as one of the four selected STS flight experiments.The purpose of the ST8 DM development was to validate, in the natural space radiation environment, the underlying fault, fault tolerance, and performance models embodying the technology (the STS flight was unfortunately canceled in August, 2007).The quality of the in-flight validation /calibration of the STS Dependable Multiprocessor Technology is a function of the number and distribution of SEUs experienced during the flight experiment. This is the critical set of parameters to be monitored because they are the basis for the predictive system reliability and availability models used to meet DM Level I requirements. The DM flight experiment system is instrumented to collect radiation event and system response data to support our validation strategy.