M. Hecht, J. Agron, H. Hecht, and K. H. Kim, "A distributed fault tolerant architecture for nuclear reactor and other critical process control applications," in Digest of the 21st Annual International Symposium on Fault-Tolerant Computing, (Montreal, Canada), pp. 3-- 9, June 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....integration of software and hardware fault tolerance techniques for distributed systems has not received enough attention. Several fault tolerance schemes have been proposed to deal with both software and hardware faults by mapping diverse software versions to redundant computer nodes (see [1, 2], for example) However, the scope of those efforts was limited to method development for utilizing redundant resources in a distributed system to protect the execution of diverse versions of a non distributed application software against hardware failure. Simultaneous tolerance of software and ....

M. Hecht, J. Agron, H. Hecht, and K. H. Kim, "A distributed fault tolerant architecture for nuclear reactor and other critical process control applications," in Digest of the 21st Annual International Symposium on Fault-Tolerant Computing, (Montreal, Canada), pp. 3-- 9, June 1991.

.... process elements to obtain what is called FTU (Fault Tolerant Unit) Others examples are DELTA 4 from LAAS (Arlat 1990) a distributed system with fault tolerant characteristics, the VOLTAN family of soft replicated nodes (Vilar 1995) or the EDRB (Extended Distributed Recovery Block) architecture (Hecht 1989). All these systems share with the proposed here the basic principles of operation, differentiating from each other in the communication channel or links between elements and their purpose. Thus, as aspects common to all, we can distinguish the following parts (Figure 1) There is a ....

Hecht, M. Agron, J. and Hochhauser, S. 1989. "A distributed fault tolerant architecture for nuclear reactor control and safety functions". Fault tolerant computing, IEEE.

....in a DRB station increases. For the sake of simplicity in discussion, these cases of using more than two try blocks in a DRB station will be treated as special cases throughout the remainder of this report. 3.2. 4 Supervisor station A major extension of the basic DRB scheme made by Hecht et al. [16, 15] was in incorporating a supervisor computing station into the LAN based system. A centralized form of a supervisor station was incorporated in [15] but many of the supervisor station functions can also be decentralized. Demonstrations of decentralized implementations have yet to take place. The ....

M. Hecht, J. Agron, and S. Hochhauser. A Distributed Fault Tolerant Architecture for Nuclear Reactor Control and Safety Functions. In Proceedings of the Real-Time Systems Symposium, pages 214--221. IEEE Computer Society, December 1989.

.... of the scheme in practical application contexts have been conducted [2, 10, 14, 16] A small company located in Los Angeles (SoHaR, Inc) extended the DRB scheme for use in real time local area PC networks for nuclear reactor control applications and produced a commercial product prototype [10, 11]. However, further research is needed to establish guidelines for cost effective design of recovery blocks. The DRB scheme will be discussed in more detail in Section 17.5. 17.4.2 Basic Techniques for Real Time Network Surveillance and Reconfiguration (NSR) The techniques for real time NSR aim ....

....amputation of faulty components, and (2.2) redistribution of tasks, including assignment of tasks to newly incorporated or repaired nodes. Basically, the following three types of approaches are conceivable: Sec. 17.5 The DRB Scheme for Real Time FT Computing Stations 417 (1) Centralized [9, 11, 18], 2) Decentralized [7, 8, 19, 21] and (3) Hybrid. Centralized approaches are simple and have been considered from the beginning days of distributed computing. Yet its integration with the techniques for construction of real time FT computing stations has not been fully accomplished. In Section ....

[Article contains additional citation context not shown here]

Hecht, M. et al., "A Distributed Fault Tolerant Architecture for Nuclear Reactor and Other Critical Process Control Applications," Proc. IEEE Computer Society's 21st Int'l Symp. on Fault-Tolerant Computing, June 1991, Montreal, pp. 462--469.

....software component and two processing nodes, and b) the two versions are not required to produce identical results and the second version need not be as sophisticated as the first version. Several demonstrations of the performance of the scheme in practical application contexts have been conducted [2, 10, 14, 16]. A small company located in Los Angeles (SoHaR, Inc) extended the DRB scheme for use in real time local area PC networks for nuclear reactor control applications and produced a commercial product prototype [10, 11] However, further research is needed to establish guidelines for cost effective ....

.... of the scheme in practical application contexts have been conducted [2, 10, 14, 16] A small company located in Los Angeles (SoHaR, Inc) extended the DRB scheme for use in real time local area PC networks for nuclear reactor control applications and produced a commercial product prototype [10, 11]. However, further research is needed to establish guidelines for cost effective design of recovery blocks. The DRB scheme will be discussed in more detail in Section 17.5. 17.4.2 Basic Techniques for Real Time Network Surveillance and Reconfiguration (NSR) The techniques for real time NSR aim ....

Hecht, M., Agron, J., and Hochhauser, S., "A Distributed Fault Tolerant Architecture for Nuclear Reactor Control and Safety Functions," Proc. IEEE Computer Society's 1989 Real-Time Systems Symp., Dec. 1989, pp.214--221.

....activate a set of fault tolerance mechanisms effective in a new mode of the environment and the computing resource. A specific instance of an AFT technique, the adaptive distributed recovery block (ADRB) scheme is a major extension of the basic distributed recovery block (DRB) scheme developed in [23, 26, 15, 25, 24]. The DRB scheme was adopted as the basic structure for designing fault tolerant real time DCSs because of its wide applicability and ability to handle both hardware and software faults with no loss of real time task executions. One fundamental software approach to realizing real time fault ....

.... 31 of various possible implementation structures has thus far taken place more in the context of highly parallel multicomputer networks [26, 25] Only in recent years have some concrete implementation structures and prototypes of DRB stations for use in real time LAN based systems been studied [15, 24]. In order to partially validate one version of a DRB implementation structure and identify detailed implementation issues, a simple experimental implementation of DRB stations in a small scale LAN based DCS testbed was conducted. issues are discussed in Section 3.2.7. 3.2 Basic Principles of the ....

[Article contains additional citation context not shown here]

M. Hecht, J. Agron, H. Hecht, and K.H. Kim. A Distributed Fault Tolerant Architecture for Nuclear Reactor and Other Critical Process Control Applications. In Proceedings of the Twenty-First International Symposium on Fault-Tolerant Computing, pages 462--469, Montreal, June 1991. IEEE Computer Society.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC