• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.05a
• /
• pp.235-238
• /
• 2003

Fuzzy C-Means(FCM) 알고리즘은 초기 군집 중심의 개수와 위치에 따라 군집 결과의 성능차이가 많이 나타난다. 하지만 일반적인 경우에 군집 중심의 개수는 분석가의 주관에 의해 결정되고, 임의적으로 결정되기 때문에 원래 데이터의 구조와는 무관하게 수행되어 최적화된 군집화 수행을 실행하지 못하는 경우가 발생하게 된다. 따라서 본 논문에서는 원래의 데이터의 구조에 좀더 근접한 퍼지 군집화를 수행하기 위하여 격자를 바탕으로 한 데이터의 밀도를 이용한 FCM을 제안하고, 이러한 밀도 기반 FCM에 의해 결정된 군집의 합병 기법을 제안하였다. N-차원의 데이터 공간을 N-차원의 격자로 나누고, 초기 군집 중심의 개수와 위치는 각 격자의 밀도를 바탕으로 결정된다. 초기화 이후에 각 격자 내부에서 FCM을 이용하여 군집화를 수행하고, 계속해서 이웃 격자의 군집결과에 대하여 군집간의 유사도 측도를 이용하여 군집 합병을 수행함으로써 데이터의 자연적인 구조에 근접한 군집화를 수행하였다. 제안된 군집화 합병 기법의 향상된 성능은 UCI Machine Learning Repository 데이터를 이용하여 확인하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2012.06c
• /
• pp.427-429
• /
• 2012

개인이 사용할 수 있는 스마트 기기가 다양해지면서 여러 기기로 생산된 사진 콘텐츠가 어떤 기준이나 규칙 없이 분산되어 있어 콘텐츠를 관리하고 원하는 콘텐츠를 검색하는 것이 어려워졌다. 따라서 본 논문에서는 개인 사진 콘텐츠를 효과적으로 분류하기 위하여 의미적 유사도를 기반으로 한 태그 클러스터링 기법을 제안한다. 태그들 사이의 유사도를 계산하여 서로 관련이 있다고 판단되는 태그들을 클러스터링 하는데, 태그가 같은 클러스터에 포함되어 있으면 그 태그를 가진 사진들도 유사성을 가진다고 볼 수 있으므로 개인 사진들을 의미에 따라 분류하는데 이용할 수 있다.

• Annual Conference of KIPS
• /
• 2010.11a
• /
• pp.906-908
• /
• 2010

클라우드 컴퓨팅 환경을 구축할 수 있게 하는 OpenNebula는 ONE 스케줄러를 통해 가상머신들의 라이프 사이클을 관리한다. ONE 스케줄러는 가상머신을 할당 할 때, 클러스터 노드의 물리적 자원 할당 여부만을 고려하기 때문에 가상 머신 생성 후의 부하를 예측하기 힘들다. 본 논문에서는 사용자의 이전 가상 머신 사용 패턴을 기반으로 부하 등급을 나누고 이 등급에 따라 가상머신을 동적으로 할당하는 기법을 제안한다.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.53 no.5
• /
• pp.13-18
• /
• 2016

In Long Term Evolution-Advanced (LTE-A), small cell enhancement(SCE) has been developed as a cost-effective way of supporting exponentially increasing demand of wireless data services and satisfying the user quality of service(QoS). However, there are many problems such as the transmission rate and transmission quality degradation due to the dense and irregular distribution of a large number of small cells. In this paper, we propose a clustering based interference management scheme in dense small cell network. We divide the small cells into different clusters according to the reference signal received power(RSRP) from user equipment(UE). Within a cluster, an almost blank subframe(ABS) is implemented to mitigate interference between the small cells. In addition, we apply the power control to reduce the interference between the clusters. Simulation results show that proposed scheme can improve Signal to Interference plus Noise Ratio(SINR), throughput, and spectral efficiency of small cell users. Eventually, proposed scheme can improve overall cell performance.

• The KIPS Transactions:PartA
• /
• v.15A no.4
• /
• pp.199-210
• /
• 2008

A Clustering based wireless internet proxy server needs a layer-7 load balancer with URL hashing methods to reduce the total storage space for servers. Layer-4 load balancer located in front of server cluster is to distribute client requests to the servers with the same contents at transport layer, such as TCP or UDP, without looking at the content of the request. Layer-7 load balancer located in front of server cluster is to parse client requests in application layer and distribute them to servers based on different types of request contents. Layer 7 load balancer allows servers to have different contents in an exclusive way so that it can minimize the total storage space for servers and improve overall cluster performance. However, its scalability is limited due to the high overhead of parsing requests in application layer as different from layer-4 load balancer. In order to overcome its scalability limitation, in this paper, we propose a distributed layer-7 load balancer by replacing a single layer-7 load balancer in the conventional scheme by a single layer-4 load balancer located in front of server cluster and a set of layer-7 load balancers located at server cluster. In a clustering based wireless internet proxy server, we implemented the conventional scheme by using KTCPVS(Kernel TCP Virtual Server), a linux based layer-7 load balancer. Also, we implemented the proposed scheme by using IPVS(IP Virtual Server), a linux-based layer-4 load balancer, installing KTCPVS in each server, and making them work together. We performed experiments using 16 PCs. Experimental results show scalability and high performance of the proposed scheme, as the number of servers grows, compared to the conventional scheme.

• The KIPS Transactions:PartA
• /
• v.12A no.7 s.97
• /
• pp.627-636
• /
• 2005

This paper presents a PC cluster-based distributed video on demand (VOD) server that minimizes the load of the interconnection network by adopting channel bonding-based MVIA and the interval cache algorithm Video data is distributed to the disks of each server node of the distributed VOD server and each server node receives the data through the interconnection network and sends it to clients. The load of the interconnection network increases because of the large volume of video data transferred. We adopt two techniques to reduce the load of the interconnection network. First, an Msupporting channel bonding technique is adopted for the interconnection network. n which is a user-level communication protocol that reduces the overhead of the TCP/IP protocol in cluster systems, minimizes the time spent in communicating. We increase the bandwidth of the interconnection network using the channel bonding technique with MThe channel bonding technique expands the bandwidth by sending data concurrently through multiple network cards. Second, the interval cache reduces traffic on the interconnection network by caching the video data transferred from the remote disks in main memory Experiments using the distributed VOD server of this paper showed a maximum performance improvement of $30\%$ compared with a distributed VOD server without channel bonding-based MVIA and the interval cache, when used with a four-node PC cluster.

• Journal of Korea Spatial Information System Society
• /
• v.8 no.2 s.17
• /
• pp.75-89
• /
• 2006

LBS (Location-Based Service) systems have become a serious subject for research and development since recent rapid advances in wireless communication technologies and position measurement technologies such as global positioning system. The architecture named the GALIS (Gracefully Aging Location Information System) has been suggested which is a cluster-based distributed computing system architecture to overcome performance losses and to efficiently handle a large volume of data, at least millions. The GALIS consists of SLDS and LLDS. The SLDS manages current location information of moving objects and the LLDS manages past location information of moving objects. In this thesis, we implement a monitoring technique for the GALIS prototype, to allow dynamic load balancing among multiple computing nodes by keeping track of the load of each node in real-time during the location data management and spatio-temporal query processing. We also propose a buffering technique which efficiently manages the query results having overlapped query regions to improve query processing performance of the GALIS. The proposed scheme reduces query processing time by eliminating unnecessary query execution on the overlapped regions with the previous queries.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.574-576
• /
• 2021

It is necessary to introduce micro-service techniques when constructing large-scale operating systems or systems that take into account scalability. Kafka is a message queue with the pub/sub model, which has features that are well applied to distributed environments and is also suitable for microservices in that it can utilize various data sources. In this paper, we propose a data sharing method for educational video sharing services using Apache's Kafka. The proposed system builds a Kafka cluster for the educational video sharing service to share various data, and also uses a spark cluster to link with recommendation systems based on similarities in educational videos. We also present a way to share various data sources, such as files, various DBMS, etc.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.3
• /
• pp.63-67
• /
• 2024

Recently, artificial intelligence (AI) workloads encompassing various industries such as smart logistics, FinTech, and entertainment are being executed on the cloud. In this paper, we address the scheduling issues of various AI workloads on a multi-tenant cloud system composed of heterogeneous GPU clusters. Traditional scheduling decreases GPU utilization in such environments, degrading system performance significantly. To resolve these issues, we present a new scheduling approach utilizing genetic algorithm-based optimization techniques, implemented within a process-based event simulation framework. Trace driven simulations with diverse AI workload traces collected from Alibaba's MLaaS cluster demonstrate that the proposed scheduling improves GPU utilization compared to conventional scheduling significantly.

• Convergence Security Journal
• /
• v.14 no.4
• /
• pp.3-8
• /
• 2014

Wireless Ad hoc Network is threatened from many types of attacks because of its open structure, dynamic topology and the absence of infrastructure. Attacks by malicious nodes inside the network destroy communication path and discard packet. The damage is quite large and detecting attacks are difficult. In this paper, we proposed attack detection technique using secure authentication infrastructure for efficient detection and prevention of internal attack nodes. Cluster structure is used in the proposed method so that each nodes act as a certificate authority and the public key is issued in cluster head through trust evaluation of nodes. Symmetric Key is shared for integrity of data between the nodes and the structure which adds authentication message to the RREQ packet is used. ns-2 simulator is used to evaluate performance of proposed method and excellent performance can be performed through the experiment.