一种容错的空间计算机体系结构

A Fault-tolerant Space Computer System Architecture

当今空间计算机必须具有强实时下的高速处理能力和自主工作方式下的高可靠性，而对长寿命卫星而言，其可靠性要求使得任何一种模式的单机结构都难以胜任，于是各种各样的冗余方案溶进了星载计算机设计中，而有目的地识别和选择一种结构使其在有限资源的条件下最大限度地实现容错，同时又能达到所要求的性能，这正是本文所追求的目标，这里阐明的是一种模块化的容错结构，它使用简易的冗余内总线．将不同功能的冗余模块紧密地耦合在一起，从而使系统级的性能可以很方便的进行扩展，功能上可以灵活地实现集中或分布，从而达到了既适应空间计算和控制要求，又满足容错的性能要求的目标。

Nowadays, space computers must have the high speed data processing ability with strong real time and the high reliability in autonomy mode. Speaking of the long - life sateliites, the high reliability requirement make it impossible to implement for any kind of single computer architecture. So, various fault - tolerant computer architecture appears in the onboard computer system design. This paper focus on the objective of how to determinately select and identify an architecture, which can ultimately implement fault - tolerant under limited resource and condition and fulfill the requisite function as well. It describes a modular fault tolerant architecture, implemented by a simple redundant local bus tightly coupling with some redundantmodels if different functions. With this architecture,we can easily expande our system performance and realize centralizing or distributing function simply. so thesystem can meet the requisite objective of both spacedata management and control and fault-tolerant specification