抗辐射容错计算机的设计与实现

发布时间：2019-06-29 07:06

【摘要】：在航空航天领域，计算机系统常常受到辐射的影响。尤其是电磁脉冲，，给计算机系统的可靠工作带来了极大的威胁。由于计算机工作环境恶劣难于进行人工的维护，但又必须保证计算机能够安全稳定的工作，所以如何提高计算机系统的抗辐射能力，具有重要的现实意义和经济价值。传统的抗辐射技术是在机器层采用无缝机壳屏蔽技术和在部件层使用加固芯片。但是这仍然不能完全屏蔽掉电磁脉冲，耦合进入到机箱中的电磁脉冲依然会导致硬件失效。如何从计算机体系结构的层面提高计算机系统抗辐射能力成为计算机抗辐射研究的重要内容。本文首先对辐射效应和抗辐射技术进行了研究。引入容错理论，从计算机的设计角度提出了一种抗辐射容错计算机的设计方案。系统采用一个备份源的机器级冷备份待命储备系统方式实现。基于冷备份比热备份对辐射尤其是电磁脉冲有着更好的抗干扰能力，在备份原状态的选择、供电电路的设计及检测器的设置上都做出相应的考虑。然后从CPLD逻辑设计及硬件电路两方面对原理样机进行详细设计与实现。原理样机由2个主机板和一个公共板组成，每个主机板上有一个容错管理模块。容错管理模块针对辐射引发故障频繁的处理器和I/O接口电路进行故障检测，并采用瞬时故障本机复位及永久故障系统切换的故障处理机制。为验证系统故障检测和故障处理的容错机制，本文开发了一个植入式故障注入系统。注入系统可以向用以进行处理器和I/O接口故障检测的数据线注入永久和瞬时故障，模拟计算机在现实运行环境下发生的各种常见故障。最后通过故障注入系统，对原理样机的容错机制进行了故障注入实验，验证了系统故障检测机制和故障处理能力。
[Abstract]:In the field of aerospace, computer systems are often affected by radiation. In particular, electromagnetic pulse brings great threat to the reliable work of computer system. Because the computer working environment is difficult to maintain manually, but it is necessary to ensure that the computer can work safely and stably, so how to improve the radiation resistance of the computer system is of great practical significance and economic value. The traditional anti-radiation technology is to use seamless housing shielding technology in the machine layer and reinforcement chip in the component layer. However, the electromagnetic pulse can not be completely shielded, and the electromagnetic pulse coupled into the chassis will still lead to hardware failure. How to improve the radiation resistance of computer system from the level of computer architecture has become an important part of computer radiation resistance research. In this paper, the radiation effect and anti-radiation technology are studied at first. Based on the theory of fault tolerance, a design scheme of anti-radiation fault-tolerant computer is proposed from the point of view of computer design. The system is realized by a machine level cold backup standby reserve system with backup source. Based on the fact that cold backup has better anti-interference ability to radiation, especially electromagnetic pulse, the selection of original state of backup, the design of power supply circuit and the setting of detector are considered accordingly. Then the prototype is designed and implemented in detail from two aspects of CPLD logic design and hardware circuit. The prototype consists of two motherboards and one common board, each of which has a fault-tolerant management module. The fault-tolerant management module detects the fault of the processor and I / O interface circuit with frequent radiation-induced faults, and adopts the fault processing mechanism of instantaneous fault local reset and permanent fault system switching. In order to verify the fault-tolerant mechanism of fault detection and fault processing, an implantable fault injection system is developed in this paper. The injection system can inject permanent and instantaneous faults into the data lines used for processor and I / O interface fault detection, and simulate all kinds of common faults of the computer in the real running environment. Finally, the fault injection experiment of the fault tolerant mechanism of the prototype is carried out through the fault injection system, and the fault detection mechanism and fault processing ability of the system are verified.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP302.8

【参考文献】