CSHFt: A Composite Fault-Tolerant Architecture and Self-Adaptable Hierarchical Fault-Tolerant Strategy for Satellite System

Nowadays, building parallel system with high-performance commercial off-the-shelf (COTS) chips becomes the main way to improve satellite system performance greatly. System reliability is a tough issue which needs to be solved by more effective fault-tolerant scheme. Centralized fault-tolerant scheme has the risk of single point of failure (SPOF), while distributed fault-tolerant scheme is much complex and introduces large overhead. Both of these traditional methods have their own drawbacks. This paper proposes a composite self-adaptable hierarchical fault-tolerant (CSHFt) scheme which effectively integrates and expands the ideas of centralized and distributed fault-tolerant methods. It constructs a composite and symmetrical system architecture supporting both fault-tolerant methods simultaneously and realizes the self-adaptable hierarchical fault-tolerant strategy. CSHFt scheme executes system fault tolerance in the sequence of 'first centralized then distributed'. System switches its fault-tolerant mode actively according to its performing history. By combining and completing two traditional methods, CSHFt scheme eliminates the risk of SPOF, reduces the overall complexity and overhead of system fault tolerance and enhances system's real-time feature. System failure only occurs when all the system nodes are broken, which makes system highly reliable. Based on the prototype system, we verify the practical CSHFt scheme. Its performance is also analyzed and evaluated.