• International Journal of Contents
• /
• v.14 no.4
• /
• pp.30-38
• /
• 2018

Recently, HTTP Adaptive Streaming(HAS), a video streaming service over the HTTP based web platform has become common. The use of HAS service in mobile communication devices such as mobile phones and tablet PCs is rapidly expanding. This paper addresses ways to improve the quality of HAS service by enhancing the terms of viewer satisfaction. HAS systems have several internal operational processes, which can affect viewer satisfaction. Such processes include, the quality determination for the next video chunk, the TCP connections-setup procedure and the congestion control operation of the TCP. This paper proposes a transmission scheme to improve the HAS quality services over mobile web. The proposed scheme takes into consideration the past implicit communication state of the receiver's playback buffer occupancy. The results of these experiments indicate that the proposed scheme can improve the quality of HAS service from the mobile viewer's point of view.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.26 no.5
• /
• pp.1141-1149
• /
• 2016

Deduplicated filesystem can reduce usage of storage. However, it be able to recover deleted block. We studied sanitization of deduplicated filesystem, LessFS which is based on FUSE(Filesystem in USErspace). First, we show a vulnerability recover deleted data in the deduplicated filesystem. We implement sanitization of deduplicated filesystem considering the part of fingerprint DB with data blocks. It takes 60~70 times compared to without sanitization. Which means access time to fingerprint DB and overhead derived from increase of number of chunk have a critical impact on sanitization time. But in case of more than 65,536 Byte of chunksize, it is faster than normal filesystem without deduplication.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.27 no.6
• /
• pp.1373-1383
• /
• 2017

Deduplication is a function to effectively manage data and improve the efficiency of storage space. When the deduplication is applied to the system, it makes it possible to efficiently use the storage space by dividing the stored file into chunks and storing only unique chunk. However, the commercial digital forensic tool do not support the file system analysis, and the original file extracted by the tool can not be executed or opened. Therefore, in this paper, we analyze the process of generating chunks of data for a Windows Server 2012 system that can apply deduplication, and the structure of the resulting file(Chunk Storage). We also analyzed the case where chunks that are not covered in the previous study are compressed. Based on these results, we propose the method to collect deduplicated data and reconstruct the original file for digital forensic investigation.

• Journal of KIISE:Databases
• /
• v.29 no.4
• /
• pp.262-273
• /
• 2002

MOLAP is a technology that accelerates multidimensional data analysis by storing data in a multidimensional array and accessing them using their position information. Depending on a mapping scheme of a multidimensional array onto disk, the sliced of MOLAP operations such as slice and dice varies significantly. [1] proposed a MOLAP cube storage scheme that divides a cube into small chunks with equal side length, compresses sparse chunks, and stores the chunks in row-major order of their chunk indexes. This type of cube storage scheme gives a fair chance to all dimensions of the input data. Here, we developed a variant of their cube storage scheme by placing chunks in a different order. Our scheme accelerates slice and dice operations by aligning chunks to physical disk block boundaries and clustering neighboring chunks. Z-indexing is used for chunk clustering. The efficiency of the proposed scheme is evaluated through experiments. We showed that the proposed scheme is efficient for 3~5 dimensional cubes that are frequently used to analyze business data.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.33 no.8B
• /
• pp.699-706
• /
• 2008

This paper compares throughput performance of TCP and SCTP in a variety of network environments. For experiments, we construct a Linux-based testbed and consider a set of performance metrics such as MSS(Maximum Segment Size), transmission delay, and packet loss rate. In addition, we analyze the effect of SCTP multi-streaming on throughput. From the experimental results, we can see that SCTP provides throughput gain of approximately $20%{\sim}50%$ over TCP. This performance gain comes from the distinctive features of SCTP such as chunk bundling, initial congestion window of 2 MTU and SACK(Selective ACK) based error control. In the lossy networks, we can see that SCTP multi-streaming transmissions can effectively overcome the so-called HoLB(Head-of-Line Blocking) phenomenon of TCP.

• Journal of KIISE:Software and Applications
• /
• v.36 no.4
• /
• pp.320-328
• /
• 2009

In general, there are two types of noun phrases(NP): Base Noun Phrase(BNP), and Maximal-Length Noun Phrase(MNP). MNP identification can largely reduce the complexity of full parsing, help analyze the general structure of complex sentences, and provide important clues for detecting main predicates in Chinese sentences. In this paper, we propose a 2-phase hybrid approach for MNP identification which adopts salient features such as expanded chunks and classified punctuations to improve performance. Experimental result shows a high quality performance of 89.66% in $F_1$-measure.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.20 no.2
• /
• pp.3-14
• /
• 2010

In this paper, we present the preimage attacks on step-reduced ARIRANG, HAS-160, and PKC98-Hash. We applied Aoki and Sasaki's chunk serach method which they have used in the attack on SHA-0 and SHA-1. Our attacks find the preimages of 35-step ARIRANG, 65-step HAS-160, and 80-step PKC98-Hash. Our results are the best preimage attacks for ARIRANG and HAS-160, and the first preimage attack for PKC98-Hash faster than exhaustive search.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.2
• /
• pp.87-92
• /
• 2024

In this paper, a RAG system was implemented using an existing Large Language Model (LLM) and Langchain library to develop a dental consultation chatbot. For this purpose, we collected contents from the webpage bulletin boards of domestic dental university hospitals and constructed consultation data with the advice and supervision of dental specialists. In order to divide the input consultation data into appropriate sizes, the chunk size and the size of the overlapping text in each chunk were set to 1001 and 100, respectively. As a result of the simulation, the Retrieval Augmented LLM searched for and output the consultation content that was most similar to the user input. It was confirmed that the accessibility of dental consultation and the accuracy of consultation content could be improved through the built chatbot.

• Communications of the Korean Mathematical Society
• /
• v.22 no.4
• /
• pp.587-596
• /
• 2007

In this paper, we study two-phase flow models. The chunk mix model of the two-phase flow equations is analyzed by a characteristic analysis. The model discussed herein has real characteristic values for all physically acceptable states and except for a set of measure zero has a complete set of characteristic vectors in state space.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2024.05a
• /
• pp.792-793
• /
• 2024

RAG는 정보 검색과 셍성 모델을 결합하여 주어진 주제나 질문에 관련된 지식을 생성하는 방법이다. 본 연구는 RAG의 성능을 높이기 위해 문서 내 문장의 평균 길이에 따른 청크의 크기와 오버랩 크기를 비교하여 최적화한다. 이를 통해 참조 문서의 특징에 맞춘 RAG를 개발할 수 있고, 다양한 종류의 글에 대해 맞춤형 답변을 제공할 수 있을 것으로 예상된다.