• Proceedings of the Korea Information Processing Society Conference
• /
• 2011.11a
• /
• pp.599-601
• /
• 2011

무선 센서 네트워크를 구성하는 센서노드들이 클러스터를 구성하고 선출된 클러스터 헤드가 클러스터 내의 센서노드들로부터 데이터를 받아서 병합한 다음, 기지국으로 데이터를 전달하는 클러스터 기반 라우팅 방법이 연구되어 왔다. 이 클러스터 기반 라우팅 방법에서 클러스터 헤드에 고장이 발생한다면, 해당 클러스터의 데이터는 기지국으로 전달할 수 없어 데이터 신뢰성에 문제가 생긴다. 이러한 문제를 해결하기 위해, 본 논문에서는 고장감내를 지원하는 클러스터 기반 라우팅 방법을 제안한다. 제안한 방법은 각 클러스터마다 백업 클러스터 헤드를 지정하여 원래의 클러스터 헤드에 고장이 발생한다면 백업 클러스터 헤드가 그 역할을 대신하도록 함으로써 데이터 전달의 신뢰성을 보장한다.

• 한국IT서비스학회:학술대회논문집
• /
• 2009.11a
• /
• pp.539-543
• /
• 2009

실시간으로 데이터를 처리하여 빠른 서비스를 제공하기 위해 PC 클러스터가 널리 사용되고 있다. 본 논문에서는 PC 클러스터의 한 종류인 HPC 클러스터에 전용 노드를 추가하는 대신, 방화벽 외부의 네트워크 상에 존재하는 비전용 노드의 유휴시간을 활용하도록 클러스터를 확장하여 성능을 향상시키는 경우 발생하는 NFS 등의 보안 문제를 SSH 터널링을 사용하여 해결하는 방안을 제시하고 암호화된 NFS의 성능을 실험하였다.

• Journal of the Economic Geographical Society of Korea
• /
• v.6 no.2
• /
• pp.327-338
• /
• 2003

The concept of industrial cluster began with the agglomeration economies occurring in a small scale place, which are mainly drawn from intense networking between companies, universities, research centers and other institutional bodies. Recently, however, as cooperation between autonomous administrative units and between industrial clusters has been more active than ever, there has emerged the concept of extended industrial cluster (EIC) and, subsequently, policy measures for nurturing it. This study focuses on the concept of EIC, mainly in the perspective of the need to adopt it into the industrial cluster literature. The study identifies the raison d'etre of EIC in the respects of the flexibility of geographic range, complementarity and cooperation between functions and regions, scale economies and productivity, and the development stage of industrial clusters. It also suggests how to establish the concept of EIC in the Korean context and how to apply EIC policy in the Korean industrial policy.

• Journal of the Korean association of regional geographers
• /
• v.22 no.2
• /
• pp.397-410
• /
• 2016

Although industrial complexes have played as an engine of the Korean economy for the last 40 years, the majority of industrial complexes shows limitations to the continuous growth such as a lack of innovation capabilities and social capital, conceived as a key to transforming into clusters of innovation. To overcome those problems, the Korean government embarked on the cluster policy from the mid 2000's, focusing on promoting the endogenous development capabilities of individual industrial complexes. Drawing upon the in-depth case study of the Gumi IT cluster, one of the representative large-scale industrial complexes in Korea, the authors conclude that the cluster policy has contributed to making the Gumi IT cluster enhance the capabilities of endogenous development through the facilitation of self-organizing learning communities within the cluster.

• Cartoon and Animation Studies
• /
• s.28
• /
• pp.209-233
• /
• 2012

The animation cluster in Korea has begun as a part of the local cultural cluster in the late 1990s with components of companies, local governments, educational institutions, and human resources, which were necessary to run the cluster. And, the animation cluster was initiated for the purpose of encouraging regional economies, but the basic unit was the local on a small scale. Because of this inherent weakness in the local cluster, it was needed to add some additional strategies that could expand the local animation industry into the formal leading industry. That is why the development policy was set up, and the local promotion agency based on this policy was established. It has been several years to manage the local promotion agency, and it is reported that there have been some visible effects. But, it is found out that analyzing the outcome of small unit cluster on the basis of existing criteria on a large scale is not reliable, which means it is not possible to evaluate the outcome of local cluster in a material way. Some examples of problems are as follows. First, the local cluster was made not autonomously but deliberately. Second, the animation cluster of each province has the same market as its target. Therefore, this research - on the basis of the diamond model - focuses on analyzing the existence and role of local promotion agencies rather than evaluating the outcome itself. Through the cases of two provincial cities, Chuncheon and Bucheon, this research examines if it is possible to evaluate the outcome of local clusters managed by promotion agencies.

• Journal of KIISE:Software and Applications
• /
• v.30 no.7_8
• /
• pp.649-658
• /
• 2003

We propose a hierarchical architecture of ART2 Network for performance improvement and fast pattern classification model using fitness selection. This hierarchical network creates coarse clusters as first ART2 network layer by unsupervised learning, then creates fine clusters of the each first layer as second network layer by supervised learning. First, it compares input pattern with each clusters of first layer and select candidate clusters by fitness measure. We design a optimized fitness function for pruning clusters by measuring relative distance ratio between a input pattern and clusters. This makes it possible to improve speed and accuracy. Next, it compares input pattern with each clusters connected with selected clusters and finds winner cluster. Finally it classifies the pattern by a label of the winner cluster. Results of our experiments show that the proposed method is more accurate and fast than other approaches.

• Journal of Internet Computing and Services
• /
• v.5 no.6
• /
• pp.21-30
• /
• 2004

In Ad-Hoc network, the cluster structure enables effective use of multiple channels, reduces the number of control messages, and increase the scalability of network, Also, it is employed for reducing the number of broadcast messages in an Ad-Hoc network. With the consideration of these advantages, it is desirable that a cluster structure keeps a few clusters in the network, Generally, the cluster formation scheme based on connectivity yields fewer clusters than the other schemes. However, the connectivity based scheme may yield even more clusters than the other schemes according to the network topology. In this paper, a cluster formation scheme dividing the cluster formation into two phases is proposed. In the first phase, the lowest connectivity host in neighborhood initiates the cluster formation. Then, an adjustment procedure for affiliating a lot of the lowest connectivity hosts is employed. In the second phase, the hosts which were not affiliated to the first phase clusters are grouped into one or more clusters through criterions of connectivity and host ID. As a result, the proposed scheme yields a fewer clusters compared with existing other schemes in fully distributed method. The simulation results proves that our scheme is better than LIDCP(3) and HCCP(3).

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.12
• /
• pp.1558-1564
• /
• 2019

Corporate business clusters create a borderless business network through value chains and value creation activities. Clusters in the IT sector are crossing regional boundaries because of their rapid technological development and short product life cycle. To investigate this phenomenon, we examined the value activities of IT companies. As a result, the global expansion of the case companies was limited, but they had business clusters that exceeded the regional boundaries. However, because the order of business clusters is not advanced, value activities are focused on productivity rather than innovation. Given the rapidly evolving nature of IT, it is necessary to evolve into a business cluster that can create new value. Companies must design and implement value-creating processes to develop new technologies or enter new markets, which can lead to cluster growth. Companies must design and implement value-creating processes to develop new technologies or enter new markets, which can lead to cluster growth.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2012.05a
• /
• pp.241-244
• /
• 2012

In mobile ad-hoc networks, it should be the most important issue to reduce the power consumption of communications, because the resource of a node is limited in these networks. In existing cluster-based routing protocols, cluster headers participate in almost all communication processes. Therefore the lifetime of the cluster header is shortened and it does not have the effective route. In the proposed cluster-based routing protocol, the cluster header transmits a control packet which gives the route information to member nodes. This makes that the cluster header decreases the number of participating in communications, and that node members do not have to communicate trough the cluster header. This results in extending the lifetime of the cluster header, and having the effective route, data transmission rate and improved stability of routes.

• Journal of KIISE:Computing Practices and Letters
• /
• v.12 no.1
• /
• pp.21-30
• /
• 2006

As the amount of information to be processed by computers has recently been increased, there has been cluster computing systems developed by connecting PCs and workstations using high-speed networks. hs a result, the studies on a cluster-based DBMS have been done with a wide range and it is necessary to develop a cluster manager for monitoring and managing cluster-based DBMSs. But, a cluster manager for efficiently managing cluster-based DBMSs has been studied in a first step. In this paper, we design and implement a high-availability cluster manager for a cluster-based DBMS. It monitors the status of nodes in cluster systems as well as the status of DBMS instances in a node, by using a graphic user interface (GUI). Our cluster manager supports the high-availability for a cluster-based DBMS by perceiving errors and recovering them, based on the monitored information on the status of a server.