Minoru Koizumi

A Fault-tolerant Decentralized Control System Ensured by Autonomous Controllability

Abstract A fault-tolerant decentralized control system ensured by autonomous controllability is presented. The system satisfies the requirements of large scale systems which are fault-tolerance, on-line maintainability and on-line expandability. In the autonomous decentralized system, system totality is not predetermined and a system should be constructed by integrating autonomous subsystems. To attain autonomy of each subsystem, autonomous controllability has been defined as the property whereby each subsystem can control its own responsible control plant even if some other subsystems fail, are under maintenance or are being expanded. Necessary and sufficient condition for autonomous controllability is derived by examining sensor malfunction as opposed to controller failure. A weak coupling method and adaptive reconfiguration method for designing an autonomously controllable system are proposed. This system is apllied to a loop computer communication network system and the communication controller of each subsystem is designed to attain fault-tolerance of the system.

DCNP: data communication management system for network node processors in a distributed processing environment

Network node processors (NNPs), which can execute application programs and can cooperate with each other via high-speed LANs, have become essential components for distributed online systems. DCNP is a data communication management system for NNP. DCNP provides unified communication interfaces and high-performance and highly reliable communication services. An addressing technique which unifies program interfaces for node-to-terminal communications and internode communications is presented. Socket interface extensions for improving performance and techniques for highly reliable internode multicast communications on multiple LANs are also presented. DCNP has been implemented with a Unix kernel on fault-tolerant minicomputers and is operating in online systems.<<ETX>>

Synchronous distributed wireless network emulator for high-speed mobility: Implementation and evaluation

Wireless network emulators provide various mobile node dynamic conditions and wireless links for real implementation of protocols to be evaluated based on user-defined mobility scenarios. The target of our proposed emulator is IP-based protocols for large-scale mobile wireless networks composed of a large number of high-speed vehicles. The requirements imposed by our target are high scalability and high-speed mobility emulation while maintaining high accuracy for many hours. To satisfy this, we first introduce a distributed architecture then propose an approach called “time-based wireless link emulation.” In this approach, the mobility scenario is converted into a wireless link schedule, and each distributed processor emulates a wireless link based only on the schedule and its clock, which is precisely synchronized with the other processor clocks by using IEEE 1588 PTPv2. Intensive experiments on the performance of the latest version of our prototype emulator show that (i) high-speed mobility emulation up to 200 m/sec with 0.1 M accuracy, (ii) packet delay emulation with 100 μsec accuracy, and (iii) time-stamps based measurement of elapsed time spent by test nodes with 100 μsec accuracy have been achieved. Our proposed emulator functioned well in an experiment with a simple protocol implementation for car-to-car/infrastructure communication.

Autonomous Coordinability of Decentralized System Considering Subsystems Failures

Distribution of intelligence caused by recent rapid development of microelectronics has brought a new coordination problem which the conventional coordination technique can not solve. In order to give a solution, an autonomous coordinability is proposed as the property to satisfy that functioning subsystems can coordinate their own objectives with each other even if some other subsystems fail, are under maintenance or during expansion. A sufficient condition for autonomous coordinability is derived. Moreover, the coordination problem among computers connected by a loop computer communication network system is discussed. In the loop computer communication network system, computers have to coordinate with each other in order to adjust the amount of messages within the capacity of buffers of the network system and to prevent the overflow of messages even in the case of some partial computers’ failures. The autonomous coordination technique is applied to this problem and the coordination rule for each computer system is designed.

Symbiotic Autonomous Decentralized System on Cooperative Distributed Scheduling

Currently, System of Systems approach becomes more and more important in the field of social infrastructure to adapt various user needs flexibly and efficiently. Symbiotic autonomous decentralized systems is one of System of Systems. In this system, optimizations of both entire system and each system are expected to be achieved simultaneously. However, the symbiotic operation based on free will of each system doesnt go well, because each system has its own key performance indicator and tries to pursue the indicator. In order to promote symbiosis in the systems, we developed Cooperative Distributed Scheduling. As a result of verification with a simulator, we confirmed the total key performance indicator of each system improved by about 17%.