Performance Evaluation for a Distributed On-Board Computer

Abstract

Autonomous spacecraft, satellite communications, and Earth observation are examples of the growing significance of onboard applications with high demands for the processing power of onboard computers in the space domain. The space-grade hardware used for onboard computers does not cope with the computational demands of the current space applications. As a result, engineers are increasingly using Commercial off-the-shelf (COTS) components to build spacecraft. However, the COTS are vulnerable to cosmic radiation. The German Aerospace Center (DLR) addressed this problem and found a solution by creating a distributed system, the Scalable On-board computing for Space Avionics (ScOSA) with a heterogeneous number of nodes. This thesis focuses on selecting and applying a suitable benchmark for performance evaluation of the ScOSA distributed system. The results of the selected benchmark (OBPMark-Image Processing) show an increase in performance from 0.968 megapixels/s in one node to 1.538 megapixels/s while using three nodes.