• Journal of KIISE:Information Networking
• /
• v.28 no.1
• /
• pp.71-81
• /
• 2001

최근 병렬처리 시스템에 대한 연구는 마이크로 프로세서 제작 기술과 네트워크 기술이 발달함에 따라 고성능의 PC와 워크스테이션 여러대를 고속의 네트워크로 연결하여 구축하는 클러스터링 환경에 관심이 고조되고 있다. 그런데 이러한 클러스터링 시스템의 성능은 수행되는 애플리케이션 프로그램의 병렬성이나 xdhtls 빈도 등의 특성에 따라 달라진다. 그러므로 클러스터링 시스템의 성능을 향상시키기 위해서는 애플리케이션의 이러한 특성을 고려하여 최상의 효과를 얻기 위한 조정작업이 필요하며 그 방법중의 하나가 시스템 상에서 발생하는 통신들에 대하 스케줄링을 수행하는 것이다. 본 논문에서는 Myrinet 스위치를 사용하여 선형으로 구성한 클러스터링 시스테과 2차원 메쉬 형태로 구성한 클러스터링 시스템의 두 가지 모델을 가정하고 이들 모델 상에서 특성 시간에 주어지는 통신요청들에 대하여 순위차원 라우팅을 사용하여 메시지들을 최단시간에 전송할 수 있는 최적 통신스케줄링 알고리즘을 제안한다. 시스템 상에서 같은 방향으로 동시에 링크를 공유한는 통신들의 개수의 최대값을 $L_{MAX}$로 시스템에서 하나의 메시지가 전달되는데 걸리는 시간을 T로 정의하면 알고리즘에 의해 선형 네트워크에서의 통신 요청 집합에 대한 메시지 전송 완료 시간은 최대 $L_{MAX}{\cdot}T$, 메쉬 네트워크에서의 통신요청 집합에 대한 메시지 전송 완료시간은 최대 $frac{3}{2}{\cdot}T$임을 증명하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.10d
• /
• pp.274-278
• /
• 2006

무선 네트워크 기술의 발전으로 인해 매우 상이한 특성을 가지는 다양한 무선 시스템이 등장하였으며, 이들 무선 시스템은 상호 중첩된 영역을 가지기도 하면서 공존할 것으로 예상된다. 한편, 두 개 이상의 에어 인터페이스를 갖추고 다양한 서비스를 동시에 받을 수 있는 이동 단말들도 등장하고 있다. 이와 같은 환경에서 이동 사용자는 로밍하면서 끊김없는 통신을 지속할 뿐만 아니라 여러 개의 공존하는 무선 엑세스 시스템을 효과적으로 활용함으로써 응용의 서비스품질 요구에 대한 만족도를 높일 수 있다. 이에 본 논문에서는 차세대 네트워크 환경에서 이동 사용자가 여러 개의 서비스를 동시에 받는 경우 각 서비스에 대하여 서비스 품질을 보장하기 위한 새로운 종단 간 이동성 지원 방안을 제안한다. 이를 위해 본 논문에서는 종단 간 이동성 관리를 수행하는 IP agent 모듈을 새로이 정의하고, 핸드오버 시, 각 응용의 특성을 고려하여 응용 별로 가장 적합한 액세스 네트워크를 선택하여 핸드오버를 수행한다. 시뮬레이션을 통하여 응용 별 특성을 고려한 종단 간 이동성을 지원하는 제안하는 방안이 기존 방안들에 비해 서비스 품질의 보장이 더 잘 이루어짐을 알 수 있었다.

• Journal of Digital Contents Society
• /
• v.15 no.5
• /
• pp.621-630
• /
• 2014

The purpose of this study is to extract metaphor characteristics of Social Network Service. Social Network Service is different from existing media only available in one-way communication, each individual expands opinion by sharing daily life and opinion directly and interpreting another user's post. This Study premise that the reason of converting from passive reader to active enunciator is the metaphor characteristics of Social Network Service by 'Source Domain' and 'Target Domain'. In addition, this study examines the meaning of structure user's text production and interpreting based triple mimesis of Paul Ricoeur. This study has significance as arguing with existing study on SNS as metonymic media and suggesting metaphor characteristics and meaning of Social Network Service.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.28 no.10B
• /
• pp.938-946
• /
• 2003

Studies of Internet algorithms or policies require experiments on the real large-scale networks. But practical problems with large real networks make them difficult. Instead many researchers use simulations on the Internet topology models. Some tried to find out abstract topological properties of Internet. And several models are proposed to reflect Internet's topological characteristics better. But few studies have been performed on how to model the evolution of Internet. We propose algorithms for modeling addition and removal of nodes and accompanied change of topologies. We analyze the topologies generated by our algorithms to observe that they obey the power-laws better than those generated by existing ones. These algorithms are also expected to be helpful in predicting future topologies of Internet.

• Journal of the Korean Society for Railway
• /
• v.20 no.3
• /
• pp.413-422
• /
• 2017

In this study we investigate the importance of the subway station using network centrality measures. For centrality measures, we have used betweenness centrality, closeness centrality, and degree centrality. A new measure called weighted betweenness centrality is proposed, that combines both traditional betweenness centrality and passenger flow between stations. Through correlation analysis and power-law analysis of passenger flow on the Seoul metropolitan subway network, we have shown that weighted betweenness centrality is a meaningful and practical measure. We have also shown that passenger flow between any two stations follows a highly skewed power-law distribution.

• Journal of the Economic Geographical Society of Korea
• /
• v.19 no.3
• /
• pp.491-511
• /
• 2016

This paper attempts to understand the nature and dynamics of global value chains in the auto industry using the OECD TiVA 2015 edition on the bilateral foreign value added in exports over the period 1995-2011 and employing the techniques of social network analysis such as the computation of network measures and visualization of value added trade flows. It is shown that there has been a tendency towards increasing production fragmentation both within and across regions. The automotive value-added network is found to have small-world properties with a hierarchical, clustered and dense structure. The differences among the US, Germany and China as major suppliers of foreign value added in global automotive value chains are remarkably revealed. Although the fragmentation of production has been developed on a global scale, a dichotomous tension between center and periphery and domestic and foreign capital lies behind it.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37C no.11
• /
• pp.1036-1044
• /
• 2012

Thanks to progressive development of IT technology, the number of intelligent devices communicating each other through an In-Vehicle Network(IVN) has been steadily increasing. It is expected that the required network bandwidth and network nodes for vehicle control in 2015 will be increased by two times and one and half times as compared to in 2010, respectively. As a result, many researchers in automotive industry has showed a significant interest on industrial Ethernets, such as EtherCAT and TTEthernet. This paper addresses an analysis on transmission delay and fault recovery performance with an EtherCAT network which is being considered as an IVN. A mathematical model based on the analysis is verified through a set of experiments using an experimental network setup.

• Journal of the Korea Society of Computer and Information
• /
• v.13 no.4
• /
• pp.31-37
• /
• 2008

In this paper, we propose a dynamic job relocation strategy that considering application's characteristics in multiple grid sites. This scheme classifies application to execute in multiple grid sites by their characteristics : computing intensive application, network intensive application. Also, it eliminates the communication between sites by allocating the network intensive application in single site, thus reducing the total job execution time. But if a number of free nodes to execute the network intensive application aren't found in single site, the proposed scheme the first allocates the network intensive application in multiple sites to minimize network latency. Then if the network intensive application being executed in multiple sites suitable free nodes are found in single site. the proposed scheme relocates the application being executed in multiple sites to another single site. This results in reducing the total job execution time. Through simulation, we show that the proposed dynamic job reallocation strategy improves the performance of Data Grid environment compared with previous strategies.

• The KIPS Transactions:PartC
• /
• v.13C no.3 s.106
• /
• pp.345-352
• /
• 2006

In the integrated wired/wireless network which consists of the Internet and moving networks, multiple wireless links are used to connect a fixed host(FH) in the Internet to a mobile host(MH) in the moving network. For use in such an environment, we propose a scheme to overcome the TCP performance degradation due to the packet losses over the wireless links without losing the end-to-end TCP semantics. The proposed scheme in each mobile router(MR) allows to obtain the information regarding packet losses over the upstream wireless links based on the received packet sequence number and the ACK number. This information is delivered to the upstream router, which enables the upstream access router(AR) or MR to quickly retransmit the lost packets. The proposed scheme has the feature to quickly recover the packet losses incurred over the upstream wireless links and the performance of the proposed scheme is evaluated through simulation. It is shown that the significant performance gain can be obtained using the proposed scheme compared with the snoop mechanism which maintains end-to-end TCP semantics and does not require any additional features at the source and/or destination nodes.

• Proceedings of the Korea Contents Association Conference
• /
• 2008.05a
• /
• pp.87-91
• /
• 2008

Data compression techniques are traditional and effective to reduce the network traffic. Generally, sensor data exhibit strong correlation in both space and time. Many algorithms have been proposed to utilize these characteristics. However, each sensor just utilizes neighboring information, since its communication range is restrained. The distribution and characteristics of whole sensor data provide other opportunities to enhance the compression technique. In this paper, we propose an orthogonal approach for compression algorithm. The base station or a super node generates useful information for compression of sensor data and broadcasts it into sensor networks. Every sensor that received the information compresses their sensor data and transmits them to the base station. We define this approach as feedback-diffusion. In order to show the superiority of our approach, we compare it with the existing aggregation algorithms in terms of the lifetime of the sensor network. As a result, our experimental results show that the whole network lifetime was prolonged by about 30%.