• International Journal of Automotive Technology
• /
• v.3 no.3
• /
• pp.89-94
• /
• 2002

This paper presents an architecture for distributed control systems and its underlying methodological framework. Ideas and concepts of distributed systems, artificial intelligence, and soft computing are merged into a unique architecture to provide cooperation, flexibility, and adaptability required by knowledge processing in intelligent control systems. The distinguished features of the architecture include a local problem solving capability to handle the specific requirements of each part of the system, an evolutionary case-based mechanism to improve performance and optimize controls, the use of linguistic variables as means for information aggregation, and fuzzy set theory to provide local control. A distributed traffic control system application is discussed to provide the details of the architecture, and to emphasize its usefulness. The performance of the distributed control system is compared with conventional control approaches under a variety of traffic situations.

• Proceedings of the Earthquake Engineering Society of Korea Conference
• /
• 2006.03a
• /
• pp.563-570
• /
• 2006

Distributed control architecture is based on sharing control and data between multiple nodes on a network Communication and task sharing can be distributed between multiple control computers. Although many communication protocols exist, such as TCP/IP and UDP, they do not have the determinism that realtime control demands. Fiber-optic reflective shared memory creates the opportunity for realtime distributed control. This architecture allows control and computational tasks to be divided between multiple systems and operate in a deterministic realtime environment. One such shared memory architecture is based on Curtiss-Wright ScramNET family of fiber-optic reflective memory. MTS has built seismic and structural control software and hardware capable of utilizing ScramNET shared memory, opening up infinite possibilities in research and new capabilities in Hybrid and Model-In-The-Loop control.

• 연구논문집
• /
• s.28
• /
• pp.239-246
• /
• 1998

Today’s information technology departments face a dilemma to create a competitive advantage for the organization by developing. deploying, and managing distributed applications that scale across the LAN, WAN, and Internet, while preserving investments in systems. Applications, information. IT organizations are today being asked to build the future, without breaking systems that maintain the current business. A possible answer to this dilemma is the implementation of a multi-tier distributed computing architecture. Multi-tier architecture has the potential to provide better, more timely information across the enterprise at a lower cost than the current combination of PC LAN, two-tier client/server, or mainframe applications that have been developed in most organizations. In this paper, we implement management information system based on distributed object middleware using multi-tier distributed computing architecture. This system uses Microsoft's implementation of DCOM/ActiveX and provides easily accessible web interface to the system users.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.11 no.5
• /
• pp.217-228
• /
• 2011

In centralize CDN systems, the content server selection is performed by service node for every user request, and the selected node is notified to the user. In this paper, we present distributed CDN architecture and algorithm in which the request from a user is delivered to the content source by a P2P algorithm utilizing DHT(distributed hash table) through the overlay network and the user selects one of the source nodes based on real-time user-centric criteria. For this purpose, we propose a modified Pastry algorithm for contents registration, search and selection, in addition to the distributed architecture. The proposed architecture has the advantages of load balancing, traffic balancing, scalability, fault-tolerance due to the self-configuration, self-healing attributes of distributed architecture. Various simulation shows the feasibility of the proposed architecture and algorithm, and the performance is compared and discussed for the variations of the proposed scheme.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.19 no.8
• /
• pp.1578-1586
• /
• 1994

In this paper, a new overload control scheme is proposed for a stored program control system. The proposed control scheme is fit for a distributed architecture. In a distributed architecture, most of the call control functions are distributed to many subscriber control processors(SP) and some functions as number translation and routing control are centralized to central processors(CP). Therefore the performance of the CP is critical to the switching system. We needed a control scheme that ensures the proper operation of the CP in overload condition. We conducted lots of simulation experiments to evaluate the performance of the proposed overload control scheme and concluded that the proposed overload control scheme is very effective for a distributed architecture.

• Journal of the Korea Society for Simulation
• /
• v.12 no.3
• /
• pp.69-81
• /
• 2003

To manage a cluster of distributed server systems, a number of management aspects should be considered in terms of configuration management, fault management, performance management, and user management. System performance monitoring and reporting take an important role for performance and fault management. In this paper, we present distributed system architecture modeling of a performance monitoring and reporting tool. Modeling architecture of four subsystems are introduced: node agent, data collection, performance management & report, and DB schema. The performance-related information collected from distributed servers are categorized into performance counters, event data for system status changes, service quality, and system configuration data. In order to analyze those performance information, we use a number of ways to evaluate data corelation. By using some results from a real site of a company and from simulation of artificial workload, we show the example of performance collection and analysis. Since our report tool detects system fault or node component failure and analyzes performances through resource usage and service quality, we are able to provide information for server load balancing, in short term view, and the cause of system faults and decision for system scale-out and scale-up, in long term view.

• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.992-995
• /
• 1996

The major roles of Power SCADA System are continuous monitoring of electrical equipments state, real-time data processing and dispatching. Especially, SCADA system demands fast response time in heavy load condition, high reliability, fault tolerance, expansion capacity for the future. According to developing computer system technology, SCADA system is changing system configuration from centralized processing system to distributed processing system. This paper describes operational benefits, problems and improvement (which is studying in theory) in the application of Open System Architecture SCADA which has been installed since 1994, Seoul regional control center in KEPCO.

• Journal of Ocean Engineering and Technology
• /
• v.22 no.3
• /
• pp.47-57
• /
• 2008

Many accidents occur on the oceans. Accidents involving ships can cause enormous damage. When an accident occurs, the cost of environmental restoration can bankrupt even a top-ranked international company. The potential damages resulting from ship accidents have resulted in many international agreements such as MARPOL, SOLAS, ISM Code, etc. The viability of damaged ships has become more and more important. A damaged ship's viability and cargo safety can be improved by a safety assessment design procedure. We consider th￡ shipbuilding industry's distributed characteristics to define a safety assessment system with software engineering-based requirements analysis. We used a distributed network-based simulation method (HLA - High level architecture), because of ease of reuse and expansion of existing components to other situations, and because it is a military standard. HLA is the standard for distributed network-based simulation in many countries including the United States, Canada, Israel, and Korea. The paper describes research to develop a prototype of a network-based safety assessment simulation system by software engineering based analysis.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.3
• /
• pp.911-931
• /
• 2021

Recently, most cloud services use Docker container environment to provide their services. However, there are no researches to evaluate the performance of communication libraries for multi-GPU based distributed deep learning in a Docker container environment. In this paper, we propose an efficient communication architecture for multi-GPU based deep learning in a Docker container environment by evaluating the performances of various communication libraries. We compare the performances of the parameter server architecture and the All-reduce architecture, which are typical distributed deep learning architectures. Further, we analyze the performances of two separate multi-GPU resource allocation policies - allocating a single GPU to each Docker container and allocating multiple GPUs to each Docker container. We also experiment with the scalability of collective communication by increasing the number of GPUs from one to four. Through experiments, we compare OpenMPI and MPICH, which are representative open source MPI libraries, and NCCL, which is NVIDIA's collective communication library for the multi-GPU setting. In the parameter server architecture, we show that using CUDA-aware OpenMPI with multi-GPU per Docker container environment reduces communication latency by up to 75%. Also, we show that using NCCL in All-reduce architecture reduces communication latency by up to 93% compared to other libraries.

• Journal of Ocean Engineering and Technology
• /
• v.17 no.5 s.54
• /
• pp.82-87
• /
• 2003

A practical method to investigate the sharing architecture of FMS, for small and medium enterprises, on distributed heterogeneous environmental conditions is investigated. Because the industrial FMS model requires a reasonable size for actual work-site flexible operation, the initial facility expense is very heavy. The support for flexible cooperation of FMS, with some external different FMS, can reduce expenditures and provide greater functionality. While most FMS provide their operation process using an adequate mechanism, when used in isolation, such as island of automation, there is increasing demand for FMS that can cooperate. By exchanging intermediate products and information with other external FMS, products can be produced that cannot be produced alone, such as those that require various processes. This study first considers the sharing architecture of FMS for small and medium enterprises, and investigates the effects of the general sharing architecture of FMS on distributed heterogeneous environments, by using transducer approach to agentification. In this study, set-up station or in-out station takes charge of the gateway of intermediate products, and routes them among distributed heterogeneous FMS plants. The transducer approach is applied to web agent and set-up agent established interfaces, with system controller and setup controller. In this study, the FMS-established sharing architecture is able to obtain not only the internal flexibility, but also the external flexibility.