• 디지털콘텐츠학회 논문지
• /
• 제6권1호
• /
• pp.85-88
• /
• 2005

분산 멀티미디어 응용에 대한 요구는 정교한 QoS 관리를 위해서 필요하다. 분산 멀티미디어 시스템 견지에서 보면, QoS의 가장 중요한 범주는 시간성, 볼륨, 신뢰성이다. 본 논문에서는 결함 허용을 통해서 신뢰성을 향상시키는 방법에 대해서 기술한다. ERA는 분산 멀티미디어에서 하나의 소프트웨어 오류를 자동적으로 복구할 수 있는 시스템이다. 본 논문은 규칙-기반 DEVS 모델링과 시뮬레이션 기법을 사용하면서 분산 멀티미디어 상에서의 오류 복구 시스템의 성능 분석을 설명한다. DEVS에서 하나의 시스템은 시간 입력, 상태, 출력 및 함수들을 가지고 있다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권9호
• /
• pp.4555-4571
• /
• 2019

Due to increasing interest in distributed agile software development, there is a need to systematically review the literature on challenges encountered in the agile software development environment. Using the Systematic Literature Review (SLR) approach, 32 relevant publications, dated between 2013 and 2018 were selected from four electronic databases. Data from these publications were extracted to identify the key challenges across the system development life cycle (SDLC) phases, which essentially are short phases in each agile-based iteration. 5 types of key challenges were identified as impacting the SDLC phases; these challenges are Communication, Coordination, Cooperation, Collaboration and Control. In the context of the SLDC phases, the Communication challenge was discussed the most often (79 times, 33%). The least discussed challenges were Cooperation and Collaboration (26 times, 11% each). The 5 challenges occur because of distances which occur in distributed environment. This SLR identified 4 types of distances which contribute to the occurrence of these key challenges - physical, temporal, social-cultural and knowledge/experience. Of the 32 publications, only 4 included research which proposed new solutions to address challenges in agile distributed software development. The authors of this article believe that the findings in this SLR are a resource for future research work to deepen the understanding of and to develop additional solutions to address the challenges in distributed agile software development.

• 한국정보처리학회논문지
• /
• 제7권4호
• /
• pp.1227-1235
• /
• 2000

In this paper, we propose new concepts that the distributed counter value with the distributed EC system identify each user who users the distributed system. The web user should register his/her own user ID in the cyber shopping mall system. Instead of registration, this paper proposes the proprietary mechanism that is distributing counter. The counter assigns the distinguished number to each client. The distributed lock algorithm is used for mutual assignment of the counter to each client. The proposed algorithm is the best solution in the distributed environment system such as cyber shopping mall. If a user should register his/her own ID in every EC system, he/she may not try to use these uncomfortable systems. The mutual counter is used to identify each client. All of these features are designed and implemented on Windows NT web server. Also these features were experiments with 5 clients for 300 times. According to the experiments, clients have their own mutual counter value. The proposed algorithm will be more efficient in internet application environment. Moreover, it will improve the number of internet users.

• 대한안전경영과학회지
• /
• 제6권3호
• /
• pp.153-160
• /
• 2004

An issue of globalized and competitive business environment is a transformation of manufacturing strategy for rapidly product development. To copy with these trends, building up a market from various product is becoming conspicuous the core strategy of manufacture. Therefore, it is increasing of necessity for the environment of collaborative product development. As a part of corresponding for these demands, we propose the method that allows easy and rapid integration of legacy resources within the company and between departments. The proposed framework can easily construct a distributed environment for collaborative design between departments in the companies. It was implemented using CO/sup TM/ based on DOME (Distributed Object-based Modeling Environment) that is knowledge based engineering solution.

• 한국CDE학회논문집
• /
• 제3권4호
• /
• pp.283-292
• /
• 1998

The engineering process including design, analysis/evaluation and manufacturing activities is becoming one of the key issues to embody a concurrent engineering concept. This paper proposes a framework to integrated the complicated engineering design and manufacturing processes under the concurrent engineering environment. The framework offers the following facilities: (1) to represent the complicated engineering process (2) to coordinate design activities and execute the process in a distributed environment (3) to support a communication among the related engineers. The engineering processes is depicted using process flow graphs that consist in tasks and the corresponding input and output data. The engineering activities in the defined processes can be executed in a distributed environment through process controller of the framework. Engineers can communicate to suggest their opinions and to exchange product information in the framework. We have conformed the CORBA standard to integrate various distributed engineering the and communicate among them, and used a Java to support the platform independent environment on the Internet. Since the proposed framework an be a formal approach to integrate the engineering processes by providing formalism, parallelism, reusability, and flexibility, it can be effectively applied to embody the concurrent engineering concept in a distributed environment.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권3호
• /
• 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.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2002년도 Proceedings of International Symposium on Remote Sensing
• /
• pp.145-150
• /
• 2002

Web GIS (Geographic Information Systems) service systems provide the various GIS services of analyzing and displaying the spatial data with friendly user-interface. These services are expanding the business domain and many users want to access the distributed various spatial data. But, it is difficult to access diverse data sources because of different spatial data format and data access methods. In this paper, we design and implement web GIS services based on the inter-operability and GML (Geography Markup Language) of OGC(Open GIS Consortium) in web distributed environment. Inter-operability provides unique accessing method to distributed data sources based on OLE DB technology of Microsoft. In addition, GML support web GIS services based on XML. We design these GIS services as components using UML (Unified Modeling Language) of an object-oriented modeling language for specifying, visualizing, constructing, and documenting the artifacts of software system. In addition, they also were developed in object-oriented computing environment, and it provides the interoperability, language-independent, easy developing environment as well as re-usability.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 컴퓨터소사이어티 추계학술대회논문집
• /
• pp.141-144
• /
• 2003

This paper proposes the distributed virtual environment framework for collaboration between heterogeneous platforms. The proposed system consists of Client, Content Server(CS), and Reflector Server(RS). The Client is composed of VR library(VRL), Interaction Manager(IM), Distributed Interaction Manager(DIM), and Network Interface(NI). The proposed Client structure can support various display environments. The XML-based IM and DIM lets the users define and add the interactions. The proposed Server structure synchronizes the distributed interaction between heterogeneous platforms and provides transparency to the users. Therefore, the proposed framework enables the users to implement virtual collaboration environment for their tasks. Moreover, it can also be used for research and development activities between a large number of the remote users. The usefulness of the framework was verified by applying it to the virtual heritage tour system.

• 전기학회논문지P
• /
• 제55권2호
• /
• pp.73-77
• /
• 2006

A server is simply a computer that is running software that enables it to serve specific requests from other computers, called 'clients.' Distributed systems are considered by some to be the 'next wave' of computing. A collection of probably heterogeneous distribution systems is transparent to the user so that the system appears as one local machine. This paper is going to search about whole of Client/Server distributed systems environment through network game. This paper presents game by one example of network game. The ladder game uses JAVA and embody to do random every time using Random function and remainder operation repeatedly. Analyze execution principle of network game through this tame and investigate about Client/Server's distributed environment through this.

• 한국정보과학회논문지:시스템및이론
• /
• 제26권10호
• /
• pp.1206-1216
• /
• 1999

이동 호스트란 시간과 공간의 제약 없이 필요한 기능을 수행하고, 또한 원하는 정보에 접근할 수 있는 휴대용 컴퓨터를 일컫는다. 이동 컴퓨팅 환경이란 이러한 이동 호스트들을 지원할 수 있도록 구성된 분산 환경이다. 따라서 이동 컴퓨팅 환경에서의 분산 알고리즘은 정적 호스트만을 고려한 기존의 분산 환경에서와는 다르게 이동 호스트의 여러 가지 특성들을 고려하여 제안되어야 한다. 즉, 이동 컴퓨팅 환경의 도래로 인하여 이동성과 휴대성, 그리고 무선 통신과 같은 특성들을 고려한 새로운 분산 상호 배제 알고리즘이 필요하다. 이제까지 이동 컴퓨팅 환경에서의 상호 배제 알고리즘은 토큰 링 구조에 기반을 두고 설계되었다. 토큰 링 구조는 이동 호스트들의 위치를 유지하기 위하여 높은 비용을 필요로 하는 단점을 가지고 있다. 본 논문에서는 균형 높이 트리(height-balanced tree)라는 새로운 모형을 제안함으로써 정적 분산 환경과 이동 분산 환경이 혼합된 환경에서 상호 배제 비용을 감소시킬 수 있는 새로운 알고리즘을 제안하며, 각 경우에 있어서의 비용을 산출하고 평가한다.Abstract The mobile host is a potable computer that carries out necessary functions and has the ability to access desirable informations without any constraints in time and space. Mobile computing environment is a distributed environment that is organized to support such mobile hosts. In that environment, distributed algorithms of which environment not only with static hosts but with mobile host's several properties should be proposed. With the emergence of mobile computing environments, a new distributed mutual exclusion method should be required to consider properties mobile computing system such as mobility, portability, and wireless communication. Until now, distributed mutual exclusion methods for mobile computing environments are designed based on a token ring structure, which have the drawbacks of requiring high costs in order to locate mobile hosts. In this paper, we propose a distributed mutual exclusion method that can reduce such costs by structuring the entire system as a height-balanced tree for static distributed networks and for networks with mobile hosts. We evaluated the operation costs in each case.