웹 트래픽 내용 분석 기반 정보보호시스템의 성능평가 방법

박웅규(Ung-Kyu Park) 한국엔터테인먼트산업학회 2015 한국엔터테인먼트산업학회논문지 Vol.9 No.1s

최근 빈번하게 발생하고 있는 악성 코드에 의한 자료 유출 침해 사고 중에 악성코드의 약 55% 이상이 웹 트래픽 기반이며 이는 계속 증가 추세에 있다. 이에 따라 웹 트래픽 내용 분석을 통하여 침해 사고를 방지하고자 하는 정보보호시스템의 개발이 필요하다. 그러나 기존의 정보보호시스템 성능평가 도구 및 방법들은 실제 상황과 유사한 웹 트래픽을 모사하는 환경을 고려하지 않아서, 새로운 정보보호시스템의 실제 운영환경에서의 성능을 신뢰성 있게 측정하는 것이 어렵다. 본 논문에서는 웹 트래픽 분석 기반 정보보호시스템의 성능평가를 위하여 실망 패킷을 재현하는 클라이언트-서버 시뮬레이션 기법을 활용한 방법을 제안한다. 그리고 본 논문에서 제안한 방법의 유용성과 타당성 검증을 위하여 실제 개발 중인 시스템의 성능평가에 적용한 사례와 결과를 기술한다. 성능평가 사례를 통하여 본 논문에서 제안한 방법이 HTTP 패킷 내용을 분석하여 자료유출을 방지하는 정보보호시스템의 실제 운영환경에서의 최대 네트워크 처리 용량 측정에 잘 적용됨을 확인하였다. Recently, information leakage due to malware has been frequently occurred. More than 55% of the information-stealing malware communications are web-based traffic and its percent is continuously increasing. Consequently, implementation of new information protection system that analyzes web traffic contents to prevent the threats is necessary. However, the previous performance evaluation methodologies are difficult to measure reliably the performance of the new system in the real operated environment because they did not consider the environment implementation to simulate similarly a real world web traffic. In this paper, we propose an adequate method using client-server simulation technology to recreate real world packets for performance evaluation of the new system. We also describe a case study of the performance evaluation applied to a new information protection system for the purpose of verification of our proposed method. The result of the case study confirms that our proposed method can be applied well to measuring the maximum network processing performance of the information protection system in the real operated environment.

정보 검색 인덱스와 병렬화 방안에 대한 연구

박웅규(Park, Ung Kyu) 서원대학교 미래창조연구원 2003 응용과학연구 Vol.12 No.1

시멘틱 웹 기술을 이용한 개념도 평가 시스템의 설계 및 구현

박웅규(Ung-Kyu Park) 한국컴퓨터정보학회 2009 韓國컴퓨터情報學會論文誌 Vol.14 No.6

최근 수십 년 동안 개념도는 의미 있는 교수 및 학습 도구로써 사용되어지고 있다. 많은 기존의 연구는 개념도가 학습자의 학습에 긍정적인 영향을 준다는 사실을 입증하였다. 그러나 기존 방법의 단점은 개념도를 평가하거나 학습자에게 피드백을 제공하는 과정이 평가자에 의해 수작업으로 이루어지도록 설계되어 있어 매우 시간 소모적인 작업이 된다는 것이다. 따라서 온라인 학습 활동을 통하여 개념도를 자동으로 처리하여 시간 복잡성을 줄이는 방법을 개발하는 것이 필요하다. 이미 기존 연구의 결과로써 오프라인 환경에서 개념도 평가를 위한 많은 점수 부여 방법이 제안되었다. 본 논문에서는 온라인 환경에 적합하도록 시멘틱 웹 기술을 적용하여 기존의 방법들을 효과적으로 구현한 자동화된 평가 시스템을 제시하였다. 본 논문을 통하여 시멘틱 웹 기술이 개념도의 효율적인 관리 및 자동화된 점수 부여를 위하여 이용될 수 있음을 확인하였다. Over recent decades, concept mapping has been used as a valuable Learning and Teaching tool. A number of studies have shown a positive impact on student learning. One of the disadvantages of this technique has been that assessing them or providing feedback to students is time consuming. We aim here to introduce ways of reducing the complexity of using concept map techniques in online activities. Several types of scoring methods for the concept map based assessment have been developed. In this paper, we describe the development of an automatic assessment system that implements those techniques. We contribute a design that uses semantic web technologies for both the management and the scoring of the concept maps.

데이터 편재 하에서 히스토그램 변환기법에 기초한 효율적인 병렬 해쉬 결합 알고리즘

박웅규(Park Ung Kyu),최황규(Choi Hwang Kyu),김탁곤(Kim Tag Gon) 한국정보처리학회 1997 정보처리학회논문지 Vol.4 No.2

In this paper, we first propose a data distribution framework to resolve load imbalance and bucket overflow in parallel hash join. Using the histogram equalization technique, the framework transforms a histogram of skewed data to the desired uniform distribution that corresponds to the relative computing power of node processors in the system. Next we propose an efficient parallel hash join algorithm for handling skewed data based on the proposed data distribution methodology. For performance comparison of our algorithm with other hash join algorithms, we perform simulation experiments and actual execution on COREDB database computer with 8-node hypercube architecture. In these experiments, skewed data distribution of the join attribute is modeled using a Zipf-like distribution. The performance studies indicate that our algorithm outperforms other algorithms in the skewed cases.

클라우드 컴퓨팅 기술 현황 분석 및 전망

박웅규(Park, Ung-Kyu) 서원대학교 미래창조연구원 2009 과학과 문화 Vol.6 No.1

With advance in the Internet and ubiquitous technology, cloud computing is becoming more and more common in an aspect of computing services. Therefore, cloud computing service industry are being promoted by activation of the global IT companies, such as Amazon, Google, and Microsoft, and considered as the key incremental service industry in the next generation. This study was performed to survey the recent status of cloud computing technology and search for development schemes of the technology to promote the industry in the future. The results of this study are as follows. First, we describe the current technology trends in the three fields such as virtualization, distributed data management, and cluster management. As follows, second, we summarize the current activities for virtualization service. Finally, we present future prospects of cloud computing and some technical research subjects for promoting the future cloud computing service.

ABRN ; 주문형 멀티미디어 데이터베이스 서비스 시스템을 위한 버퍼 교체 알고리즘

정광철(Jung Kwang Chul),박웅규(Park Ung Kyu) 한국정보처리학회 1996 정보처리학회논문지 Vol.3 No.7

In this paper, we address the problem of how to replace byffers in multimedia database systems with time-varying skewed data access. The access pattern in the multimedia database system to support audio-on-demand and video-on-demand services is generally skewed with a few popular objects. In addition the access pattern of the skewed objects has a time-varying property. In such situations, our analysis indicates that conventional LRU(Least Recently Used) and LFU(Least Frequently Used) schemes for buffer replacement are not suited. We propose a new buffer replacement algorithm(ABRN: Adaptive Buffer Replacement using Neural Network) using a neural network for multimedia database system with time-varing skewed data access. The major role of our neural network classifies multimedia objects into two classes: a hot set frequently accessed with great popularity and a cold set randomly accessed with low popularity. For the classification, the inter-arrival time values of sample objects are employed to train the neural network. Our algorithm partitions buffers into two regions to combine the best properties of LRU and LFU. One region, which contains the hotset objects, is managed by LFU replacement and the other region, which contains the cold set objects, is managed by LRU replacement. We performed simulation experiments in an actual environment with time-varying skewed data access to compare our algorithm provides better performance as compared to the other algorithm. Good performance of the neural network-based replacement scheme means that this new approach can be also suited as an alternative to the existing page replacement and prefetching algorithm in virtual memory systems.

분할된 근거리망 분산 데이타베이스 시스템에서 결합 연산 최적화

안명수(Myoung Soo Ahn),박종수(Jong Soo Park),이동면(Dong Myun Lee),박웅규(Ung Kyu Park),김명환(Myung Hwan Kim) 한국정보과학회 1992 정보과학회논문지 Vol.19 No.5

본 논문에서는 고속 근거리망에 구축된 분산 데이타베이스 시스템에서 분할된 두 릴레이션 사이의 결합 연산을 최적화하기 위한 효율적인 알고리즘을 제안한다. 제안된 알고리즘은 프래그먼트(fragment)결합 할당 방법을 사용하여 응답시간을 최소화한다. 프래그먼트결합 할당 문제와 평형된 그래프 분할 문제 사이의 유사성이 식별되었다. 두 문제가 모두 Np-hard 문제이므로 그래프 분할을 위해 잘 알려진 휴리스틱 알고리즘이 결합 할당을 위해 수정되어 응용되었다. 각 프래그먼트에 관한 의미정보와 복사본을 이용하여 불필요한 처리비용을 줄이고 병렬처리 능력을 향상시킨다. 모의실험 결과 제안된 알고리즘이 기존의 알고리즘들보다 좋은 성능을 보여주며 각 사이트의 부하도 평형됨이 밝혀졌다. We propose an efficient optimization algorithm for the joins between two fragmented relations in a distributed database system on a fast local network. The algorithm uses the fragment join assignment method to minimize the response time of a join. The similarities between the fragment join assignment problem and the balanced graph partitioning problem are identified. Since both problems are NP-hard, a well known heuristic algorithm for graph partitioning is modified to solve the fragment join assignment problem. Semantic information associated with each fragment and duplicate copies are used to eliminate unnecessary processing costs and to increase parallism. Simulation results show that the proposed algorithm outperforms the existing algorithms and that the loads of sites are balanced.

미래 e-러닝 시나리오에 기반을 둔 e-러닝 표준화 로드맵

최현종 ( Hyun Jong Choe ),조용상 ( Young Sang Cho ),박웅규 ( Ung Kyu Park ),김태영 ( Tae Young Kim ) 한국컴퓨터교육학회 2007 컴퓨터교육학회 논문지 Vol.10 No.2

연속 미디어 서비스를 지원하는 멀티미디어 서버의 구성 방법

박웅규 서원대학교 미래창조연구소 2004 과학과 문화 Vol.1 No.1

With advances in computer and communication technology, the on-line multimedia service system is becoming more and more common in various aspects of life, including education, training, entertainment, and publishing. To guarantee the quality of service in the system, it is necessary to play the continues media. such as video or audio, without stopping. However, since the traditional system do not consider the constraint, it is inadequate for a system to support the continuous media service. In this paper, we discuss technical issues for implementing a multimedia server to support efficiently the continuous media service. Firstly, we present an organization methodology of the continuous media server and discuss an application area of the server. Secondly, we present an overall software architecture for the server and discuss the implementation methodology of several software modules in the system. Finally, we present an efficient overall operating algorithm of the server to support the continuous media service.

멀티쓰레드 데이타베이스 시스템에서 효율적인 동적 메모리 관리 기법 연구

박웅규 서원대학교 응용과학연구소 2001 응용과학연구 Vol.10 No.1

With advances in parallel processing technology, multithreaded programming environment is becoming more and more common in computer application areas. Thus most commercial DBMS systems have been developed on multithreaded programming environment. However, most key technology, such as dynamic memory management methodology, is not yet open. In this paper, we propose an efficient methodology for dynamic memory management in the multithreaded database system. Firstly, we discuss several critical problems to be occurred when multithreaded database system is implemented to use the existing dynamic memory management methodology. We propose a new methodology to handle the problems and present an efficient method to implement the proposed methodology on SUN Solaris 2.5 operating system environment. Finally, we discuss efficiency and usefulness of our proposed methodology on multithreaded database system and further studies.