http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
교육 및 연구전산망 설계를 위한 분산처리 시스템에 관한 연구 : 전산망의 효과적인 설계를 중심으로
백두권,황종선,이태원 한국생산성학회 1990 生産性論集 Vol.4 No.1
This paper discusses distributed processing systems to be applied to design the Education and research Computer Network. Especially, we have analyzed performance measures of computer communication networks, have solved the capacity assignment problem approximately, and have presented two algorithms by which we could decide the efficient topology of a computer communication network. In general, computer networks can be modelled by appropriate queueing models on the basis of queueing theory. But in that case we must consider several performance measures carefully to estimate the performance of the model of a computer network. In this paper, we have analyzed some important performance measure - time delay, line cost, and routing strategy. According to queueing theory, we can get optimal solutions of the network optimization problems. However, if we are to use this method we should know a great deal of data, and so it is time-consuming to find the analytic optimal solutions. In this paper, we have introduced the inefficient index(I) to describe another method - approximate method - of solving the capacity assignment problem. Meanwhile, we have presented two algorithms by which the efficient network topology can be extracted. In other words, this paper has shown construct the Education and Research Computer Network, which is one of the Five Main National Computer Networks, and we can expect it might be applied to the implementation of other computer networks. the design method of the network topology satisfying suitable utilities and costs of processors, which are components of a computer network, under the given design constraints. With applying those algorithms to the computer network including several distributed database systems, we have presented a case study. In conclusion, this paper has provided a theoretical background to
백두권,최요한,박성공,이정욱,정동원,Baik, Doo-Kwon,Choi, Yo-Han,Park, Sung-Kong,Lee, Jeong-Oog,Jeong, Dong-Won 한국정보처리학회 2003 정보처리학회논문지D Vol.10 No.2
To share and standardize information, especially in the database environments, MDR (Metadata Registry) can be used to integrate various heterogeneous databases within a particular domain. But due to the discrepancies of data element representation between organizations, global information integration is not so easy. And users who are searching integrated information on the Web have limitation to obtain schema information for the underlying source databases. To solve those problems, in this paper, we present a 3-layered Information Integration System (LI2S) based on MDRs and Ontology. The purpose of proposed architecture is to define information integration model, which combine both of the nature of MDRs standard specification and functionality of ontology for the concept and relation. Adopting agent technology to the proposed model plays a key role to support the hierarchical and independent information integration architecture. Ontology is used as for a role of semantic network from which it extracts concept from the user query and the establishment of relationship between MDRs for the data element. (MDR and Knowledge Base are used as for the solution of discrepancies of data element representation between MDRs. Based on this architectural concept, LI2S was designed and implemented. 한 도메인 내에서 다양한 데이터베이스의 데이터를 공유하고 표준화하기 위해 MDR(Metadata Registry)을 이용하여 정보를 통합할 수 있다. 그러나 MDR을 구축하는 조직간 데이터요소 표현의 불일치 때문에 MDR간 광역적인 정보를 통합하는데 어려움이 있다. 또한 웹과 같은 다양한 데이터베이스가 존재하는 환경에서 통합된 정보를 검색하고자 하는 사용자는 각각의 데이터베이스 스키마 정보를 확보하기엔 한계가 있다. 따라서 본 논문에서는 MDR과 온톨로지(Ontology)를 결합한 3계층 정보 통합 시스템을 제안한다. MDR간 데이터요소의 관계를 사상시키고 표현의 불일치를 해결하기 위해 MDR의 표준성기능과 온톨로지의 개념과 관계기능을 결합한 정보 통합 모델을 정의하고 에이전트 기술을 적용한 계층적이고 독립적인 정보 통합 아키텍처를 제안한다. 온톨로지는 사용자의 질의에서 개념을 추출하기 위한 의미망(semantic network)의 역할과 MDR간의 데이터요소 관계를 설정하기 위한 기능으로 적용되었다. MDR과 지식베이스(Knowledge Base)는 데이터요소간 표현 불일치를 해결하기 위해 적용하였다. 이러한 핵심요소를 고려하여 제안된 아키텍처를 사용하여 MDR과 온톨로지를 결합한 3계층 정보 통합 시스템을 구현하였다.