http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
분산 병렬 환경에서 이동객체에 대한 위치기반 연속 이벤트 검출 기법
Ariunerdene Nyamdavaa,송석일(Seokil Song) 한국정보기술학회 2020 한국정보기술학회논문지 Vol.18 No.8
In this paper, we design and implement a continuous event detection method for large mobile object data streams in a distributed parallel environment. The proposed method includes a distributed in-memory grid index and a continuous query processing method based thereon for effective event detection for moving objects. The proposed method detects pre-defined events by setting a region of interest and analyzing the location data stream of moving objects in a continuous manner. Subsequently, in addition to the table storing the location data of the moving object, we propose a method of separately maintaining a table storing the moving object data in the region of interest for continuously detecting events. The proposed method is implemented based on Apache Kafka, Spark, and Hbase, and performance is evaluated through various experiments.
익스텐트 맵을 이용한 온라인 스토리지 마이그레이션 방법
김진수(Jinsu Kim),송석일(Seokil Song) 한국정보기술학회 2020 한국정보기술학회논문지 Vol.18 No.5
Online storage migration is an operation to replicate a storage remotely without interrupting service. This operation is essential for various applications such as storage replacement, replication, and data center relocation. The proposed online storage migration method in this paper creates an extent map of files to be transferred, performs sequential IO according to the physical address, reads the files, and transmits them to the target to increase the transfer speed. In addition, it is possible to track changes on the source side occurring during transmission and record them in log records to reflect the changes during transmission on the target side. Finally, in this paper, we implement the proposed online storage migration method and perform an experiment to measure the speed of online storage migration. As a result of the experiment, we show that the proposed migration method is performed twice or more faster than the simple remote copy command, scp.
하이브리드 스토리지를 위한 핫/콜드 데이터 재배치 방법
길기정(Khil Ki-jeong),송석일(Seokil Song),정승국(Seung-Kook Cheong) 한국정보기술학회 2013 한국정보기술학회논문지 Vol.11 No.1
A hybrid storage system consists of a small and fast storage (DDR-SSD) and one or more slow and large storages (HDDs) to attain high performance with low cost. In order for the hybrid storage system to yield the best attainable performance, hot data blocks of hybrid storage system should be placed on a DDR-SSD and cold data blocks should be placed on a HDD according to the temperature of each data block. In this paper, we propose a hot and cold data replacement method that measures the temperature of data on real-time and replaces the data blocks. The proposed hot and cold data replacement method includes a new temperature measure method and a data migration method based on data replication. Also, we show the superiority of our proposed method through simulation.
길기정(Ki-jeong Khil),송석일(Seokil Song),송재종(Chai-Jong Song),이석필(Seok-Pil Lee),장세진(Sei-Jin Jang),이종설(Jong-Seol Lee) 한국콘텐츠학회 2011 한국콘텐츠학회논문지 Vol.11 No.12
이 논문에서는 시퀀스 데이터베이스에서 효과적인 유사 검색을 지원하기 위한 색인 기법을 제안한다. 제안하는 색인 기법에서는 데이터 시퀀스에 대한 필터링 효과를 얻기 위해, 최소 DTW 거리를 새롭게 제안한다. 최소 DTW 거리는 유사한 데이터 시퀀스 그룹과 질의 시퀀스 사이의 최소거리를 측정하는 방법이다. 최소 DTW 거리는 계층적인 색인 구조를 통해서 시퀀스 데이터베이스를 필터링하면서 유사도 검색을 수행할 수 있도록 한다. 마지막으로, 실험을 통해서 제안하는 방법의 우수성을 입증한다. In this paper, we propose an indexing method to support efficient similarity search for sequence databases. We present a new distance measurement called minimum DTW distance to enhance the filtering effects. The minimum DTW distance is to measure the minimum distance between a sequence data and the group of similar sequences. It enables similarity search through hierarchical index structure by filtering sequence databases. Finally, we show the superiority of our method through some experiments.