• Journal of KIISE:Software and Applications
• /
• v.37 no.4
• /
• pp.311-316
• /
• 2010

Esterel is an imperative synchronous language that is used to develop memories, cache controllers, bus interfaces, and so on. An Esterel statement is called schizophrenic if it is executed more than once in an instant. A schizophrenic statement may cause problems when it is translated to hardware circuits; a circuit performs more than one reaction in a clock. Previous works claim that only local signal declarations and parallel statements may cause schizophrenic problems. However, control signals produced by a trap statement or data signals used by emit statements can cause schizophrenia. They are new schizophrenic patterns. Especially, schizophrenic problems caused by emit statements cannot be solved by a loop unrolling technique that is the key idea of previous curing techniques for schizophrenic problems. In this paper, we introduce and define the two schizophrenic problems.

• Journal of KIISE:Software and Applications
• /
• v.37 no.8
• /
• pp.647-652
• /
• 2010

Esterel is an imperative synchronous language well-adapted to control-intensive systems. When an Esterel program is translated to a circuit, the synchronizer of a parallel statement may be executed more than once in a clock; the synchronizer is called schizophrenic. Existing compilers cure the problems of schizophrenic parallel synchronizers using logic duplications. This paper proposes the conditions under which a synchronizer causes no problem in circuits when it is executed more than once in a clock. In addition we design a detection algorithm based on those conditions. Our algorithm detects schizophrenic parallel synchronizers that have to be duplicated in Esterel source codes so that compilers can save the size of synthesized circuits

• Proceedings of the Korea Multimedia Society Conference
• /
• 1998.10a
• /
• pp.196-201
• /
• 1998

최근 Mbone을 통한 다양한 회의 시스템들이 등장하고 있다. 이들 회의 시스템은 주로 영상회의 시스템과 음성회의 시스템으로 이루어져 있으며, 사람들은 영상보다 음성에 더 민감하기 때문에 음성품질이 영상품질보다 더 중요하다. 현재 인터넷은 QoS(Quality Of Service)를 제공하지 않기 때문에, 인터넷을 사용하는 음성 회의 시스템들도 근본적으로 품질을 보장할 수 없다. 따라서 음성품질을 향상시키기 위한 다양한 방법들이 연구되었고, 대표적인 기법으로 중복정보(redundant information) 복구기법이 있다. 본 논문에서는 기존의 중복정보 복구기법을 개선한 동적 중복정보 복구기법을 제안한다. 이 기법은 네트워크의 상태에 따라 중복정보의 양을 조절하기 때문에 중복정보의 효율을 향상시킬 수 있다. 실제 인터넷을 통한 실험에서 평균 9%의 성능향상이 있었다.

• Journal of KIISE:Databases
• /
• v.32 no.2
• /
• pp.177-192
• /
• 2005

In this paper, we introduce a structural redundancy method. It reduces the query processing cost incurred when reconfiguring an XML document from divided XML data in shredding XML documents into relations. The fundamental idea is that query performance can be enhanced by analyzing query patterns and replicating data essential for the query performance. For the practical and effective structural redundancy, we analyzed three types of ID, VALUE, and SUBTREE replication. In addition, if given XML data and queries are very large and complex, it can be very difficult to search optimal redundancy set. Therefore, a heuristic search method is introduced in this paper. Finally, XML query processing cost arising by employing the structural redundancy, and the efficiency of proposed search method arc analyzed experimentally It is manifest that XML read query is performed more quick]y but XML update query is performed more slowly due to the additional update consistency cost for replicas. However, experimental results showed that in-place ID replication is useful even in having excessive update cost. It was also observed that multiple-place SUBTREE replication can enhance read query performance remarkably if only update cost is not excessive.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.12
• /
• pp.907-917
• /
• 2006

In this paper, we propose a strategy to improve fault-tolerance and scalability of replicated services in mobile agent systems by applying an appropriate passive replication protocol for each replicated service according to whether the service is deterministic or non-deterministic. For this purpose, two passive replication protocols, PRPNS and PRPDS, are designed for non-deterministic and deterministic services respectively. They both allow visiting mobile agents to be forwarded to and execute their tasks on any node performing a service agent, not necessarily the primary agent. Especially, in the protocol PRPDS, after a backup service agent has received each mobile agent request and obtained its delivery sequence number from the primary service agent, the backup is responsible for processing the request and coordinating with the other replica service agents. Therefore, our strategy using the two proposed protocols can promise high scalability of replicated services a large number of mobile agents attempt to access in mobile agent systems. Our simulation results show that the proposed strategy performs much better than the one using only the traditional passive replication protocol.

• Journal of KIISE:Software and Applications
• /
• v.34 no.6
• /
• pp.503-518
• /
• 2007

Normalization transforms one parent genotype to be consistent with the other before crossover. In this paper, we explain how normalization alleviates the difficulties caused by non-synonymously redundant encodings in genetic algorithms. We define the encodings with maximally non-synonymous property and prove that the encodings induce uncorrelated search spaces. Extensive experiments for a number of problems show that normalization transforms the uncorrelated search spaces to correlated ones and leads to significant improvement in performance.

• Journal of KIISE:Computer Systems and Theory
• /
• v.37 no.1
• /
• pp.8-18
• /
• 2010

Traditional storage server suffers from duplicated data blocks which cause an waste of storage space and network bandwidth. To address this problem, various de-duplication mechanisms are proposed. Especially, lots of works are limited to backup server that exploits Contents-Defined Chunking (CDC). In backup server, duplicated blocks can be easily traced by using Anchor, therefore CDC scheme is widely used for backup server. In this paper, we propose a new de-duplication mechanism for improving a storage system. We focus on efficient algorithm for supporting general purpose de-duplication server including backup server, P2P server, and FTP server. The key idea is to adapt stride scheme on traditional fixed block duplication checking mechanism. Experimental result shows that the proposed mechanism can minimize computation time for detecting duplicated region of blocks and efficiently manage storage systems.

• Journal of KIISE
• /
• v.43 no.10
• /
• pp.1165-1172
• /
• 2016

In cloud storage environment, deduplication enables efficient use of the storage. Also, in order to save network bandwidth, cloud storage service provider has introduced client-side deduplication. Cloud storage service users want to upload encrypted data to ensure confidentiality. However, common encryption method cannot be combined with deduplication, because each user uses a different private key. Also, client-side deduplication can be vulnerable to security threats because file tag replaces the entire file. Recently, proof of ownership schemes have suggested to remedy the vulnerabilities of client-side deduplication. Nevertheless, client-side deduplication over encrypted data still causes problems in efficiency and security. In this paper, we propose a secure and practical client-side encrypted data deduplication scheme that has resilience to brute force attack and performs proof of ownership over encrypted data.

• Proceedings of the Optical Society of Korea Conference
• /
• 2002.07a
• /
• pp.18-19
• /
• 2002

1950년 Norman Ramsy는 시간적으로 분리된 두 진동 자장을 사용한 마이크로파 전이 방법을 개발했다 이 방법을 사용할 경우 그림1과 같이 넓은 선폭을 가지는 공진선 ("Rabi pedestal") 위에 좁은 선폭의 분광 신호("Ramsey fringe")를 얻을 수 있다. 램지 방법은 정밀한 분광 측정에 많이 사용되고 있으며, 특히 원자 시계를 구현하기 위한 기본 기술이 되고 있다. (중략)

• Journal of KIISE:Software and Applications
• /
• v.28 no.8
• /
• pp.593-602
• /
• 2001

G-Machine은 주어진 표현식을 그래프 축약을 통해석 계산한다. 이러한 축약 과정은 표현식을 그래프로 표현하기 위해 많은 그래프 공간을 필요로 한다. 본 논문에서는 그래프 공간을 효율적으로 사용할 수 있는 방법을 제안한다. 본 논문에서는 우회 노드 장소에 자료를 중복 사용하는 방법을 제안하여 G-Machine에서 필요로 하는 그래프 공간을 절약하며, 이를 위한 새로운 추상기계인 poGM(Pointer Overlapped GM)을 제안한다. poGM에서는 필요한 곳에만 상대 주소를 사용하여, 그래프 내부 노드와 단말 노드 모두에서 그래프 공간을 절약할 수 있다. 기억 장소 재활용 체계를 포함하는 시뮬레이션의 결과벤치마크 프로그램에서 GM보다 총합 사용량은 평균32%, 최소합은 평균 47.6% 적게 사용하였으며, 수행 시간은 평균 30% 빠르게 수행되었다.