A fault-tolerant, scalable and distributed middleware for future space missions

Abstract

The computational demands of current space missions outrun the capability of available state-of-the-art space-qualified computing hardware. Future missions, including earth-observation with high-resolution cameras, on-orbit real-time servicing, as well as autonomous spacecraft and rover missions on distant celestial bodies, will have even higher requirements concerning the computational power of the spacecrafts hardware. An approach to overcome these difficulties, which is already used and will be used widely in the future, is the use of interconnected commercial-off-the-shelf (COTS) processors for the on-board computers (OBC) of spacecraft. The COTS processors, while offering a much higher performance at a much lower cost, have the disadvantage of being more vulnerable to soft errors induced by radiation in comparison to space-qualified processors. In this context, the ScOSA Flight Experiment project develops an OBC which copes with the requirements for future space missions. The OBC will combine reliable computing nodes (RCNs) together with high-performance nodes (HPNs) into a single distributed system. For abstracting this complex architecture to a monolithic system, a middleware is needed. In this paper, we present the ScOSA middleware by means of its individual components. Furthermore, we explain its features of heterogeneity, scalability, reconfiguration, cross-processor distribution and fault-tolerance. Since the ScOSA Flight Experiment project is a successor of the OBC-NG and the ScOSA projects, its middleware is also a further development of the existing middleware. Therefore, we will present and discuss our contributions and plans for enhancement of the middleware in the course of the new project.