Network


Latest external collaboration on country level. Dive into details by clicking on the dots.

Hotspot


Dive into the research topics where Bo-kyeong Kim is active.

Publication


Featured researches published by Bo-kyeong Kim.


IEEE Transactions on Consumer Electronics | 2013

LSF: a new buffer replacement scheme for flash memory-based portable media players

Bo-kyeong Kim; Dong-Ho Lee

A portable media player is a consumer electronics device that is capable of storing and playing digital media such as audio, image, video, and etc. NAND flash memory has been widely deployed as a storage medium for these portable media players because of its small size, light weight and low power consumption. However, due to distinct characteristics of NAND flash memory such as no in-place update, asymmetric operation unit and speed, conventional disk-based buffer replacement scheme may not yield good performance on flash memory-based storage systems. In this work, we propose a novel buffer replacement scheme considering the features of metadata as well as multimedia data for portable media players. Our proposed scheme classifies the pages into a long sequential reference set and a short sequential reference set by checking the sequential continuity of pages in the buffer. And then, it mainly evicts long sequential reference sets first from the buffer and keeps short sequential reference sets in the buffer as long as possible to improve the hit ratio. By this strategy, we can efficiently reduce the number of write and erase operations to the flash memory when storing and playing digital media. Through various experiments, we show that our proposed scheme yields better performance than its related work1.


international conference on ubiquitous information management and communication | 2012

Hybrid hash index for NAND flash memory-based storage systems

Min-Hee Yoo; Bo-kyeong Kim; Dong-Ho Lee

Recently, NAND flash memory has become an alternative storage system for a storage device because of its rapid access speed, shock-resistant, small size, and lower power consumption. However, compared to the disk-based storage device, NAND flash memory has several hardware limitations such as asymmetric speed of read, write and erase operations and no in-place update. Therefore, a hash index structure, one of the popular disk-based index structures, may result in severe performance degradation if it is directly deployed on NAND flash memory. In this paper, we propose a new hash index structure that delays split operations which cause additional write and erase operations by using overflow buckets when bucket overflow occurs rather than immediate invocation of split operations. Also, we present a new overflow bucket maintenance policy that can reduce the number of additional read and write operations by adaptively handling the number of overflow buckets according to the ratio of update and delete records for each bucket. Finally, we show the superiority of the proposed index structure by comparing it to traditional extendible hash index through various experiments.


Journal of Circuits, Systems, and Computers | 2015

h-Hash: A Hash Index Structure for Flash-Based Solid State Drives

Bo-kyeong Kim; Sang-Won Lee; Dong-Ho Lee

Flash-based solid state drives (SSDs) have been widely used as storage medium because of their fast access speed, high reliability and low power consumption. In spite of these advantages, they inherit distinct characteristics such as no in-place updates, asymmetric operation speed and unit because they consist of NAND flash memories. Therefore, a disk-based hash index structure may result in severe performance degradation if it is directly deployed on a NAND flash memory-based storage system. In this paper, we propose a new hybrid hash index structure for flash-based SSDs. It delays and reduces the split operations which cause slow block erasure and additional read and write operations by exploiting overflow buckets according to the ratio of updates and deletions. Through various performance evaluations, we show the superiority of our proposed index structure by comparing it to other hash index structures for flash-based SSDs.


IEEE Transactions on Consumer Electronics | 2015

A wear-leveling algorithm exploiting k-bitwise operations for flash storage devices

Bo-kyeong Kim; Dong-Ho Lee

Flash storage devices are widely used for mobile consumer electronics due to small size, low power consumption, and high performance. Generally, the flash storage device consists of NAND flash memories. Compared to traditional magnetic disks, NAND flash memory requires an additional erase operation and its blocks have limited erase cycles. For extending its endurance, various wear-leveling algorithms have been proposed. However, they invoke many read/write/erase operations and use many memory resources for managing their block states because they do not consider the property of the flash translation layer. To solve these problems, a new wear-leveling algorithm for the log-based flash translation layer is proposed in this paper. In the log-based flash translation layer, since log blocks are frequently updated and erased, the cold block rarely removed is reserved for a next log block so that all the blocks are evenly erased. In addition, the proposed algorithm reduces the usage of memory resources by exploiting k-bitwise erase table that only needs small k-bit erase flags for managing its block erase state. Through various experiments with related wear-leveling algorithms, this paper shows the superiority of the proposed wear-leveling algorithm.


Design Automation for Embedded Systems | 2015

LSB-Tree: a log-structured B-Tree index structure for NAND flash SSDs

Bo-kyeong Kim; Dong-Ho Lee

NAND flash memory has been widely used as a storage device for embedded systems because of its fast access speed, low power consumption, and lower noise compared to a hard disk. However, due to its unique characteristics such as the lack of an in-place update and asymmetric operation speed/unit, conventional disk-based systems and applications may experience severe performance degradation when NAND flash memory is used. When a disk-based index structure such as a B-Tree is implemented in flash memory-based storage systems, intensive overwrite operations, which are caused by record insertion, deletion, and reorganization, may result in severe performance degradation. Although several index structures have been proposed to overcome this problem, they suffer from frequent node splits, rapid increments of tree height, and poor space usage. In this paper, we propose a log-structured B-Tree index structure where a log node corresponding to a leaf node is allocated for updating the modified data, and then these data in the log node are stored in a single write operation. Our proposed index structure reduces additional write operations by deferring parent node changes. In addition, the index structure reduces the number of write operations by directly switching the log node to a leaf node if the data are sequentially inserted according to key order. Through various experiments, we show that our proposed index structure performs better than related techniques.


international conference on ubiquitous information management and communication | 2013

Real-time near-duplicate web video identification by tracking and matching of spatial features

Kyung-Wook Park; Jee-Uk Heu; Bo-kyeong Kim; Dong-Ho Lee

With the exponential growth of the Web, real-time near-duplicate Web video identification is becoming more and more important due to its wide spectrum of applications including copyright detection and commercial monitoring. Though there has been significant research effort on efficiently identifying near-duplicates from large video collections, most of them use global features sensitive to photometric variations such as illumination direction, intensity, colors, and highlights. This paper proposes a novel local feature based approach in order to address the efficiency and scalability issues for near-duplicate Web video identification. Firstly, in order to represent the shot, we introduce a compact spatial signature generated with trajectories of the patches. And then, we construct an efficient data structure which indexes the spatial signatures to find the corresponding shots from query video. Finally, we adopt naive-Bayesian approach to estimate the near-duplicates from the set of corresponding shots. To demonstrate the effectiveness and efficiency of the proposed method, we evaluate its performance on an open Web video data set containing about 10K Web videos.


IEEE Transactions on Consumer Electronics | 2016

A new hash algorithm exploiting triple-state bucket directory for flash storage devices

Bo-kyeong Kim; Joo-Young Kang; Dong-Ho Lee

Recently, the flash storage device becomes more important to the mobile consumer electronics that supports various functionalities because of its portability and reliability. The capacity of the flash storage device is becoming larger and its price is going down due to the mass production of the Triple-Level Cell (TLC) flash memory. Since TLC flash memories are very cheap, they are widely deployed in the mobile flash storage device although its durability is poor. In addition, it is difficult to handle the large-scale flash memories with limited memory resources in the mobile flash storage device. Therefore, TLC-based flash storage device requires to internally optimize the write and erase operations with small memory resources. This paper proposes a new hash-based mapping algorithm, which reduces the redundant write and erase operations and uses only a few memory resources, for the mobile flash storage device. The proposed hash algorithm also guarantees uniform performance regardless of various workload patterns. Through various experiments with related works, this paper shows the superiority of the proposed hash algorithm.


computer software and applications conference | 2015

A Read-Optimized Index Structure for Distributed Log-Structured Key-Value Store

In-Su Kang; Bo-kyeong Kim; Dong-Ho Lee

Recently, Big Data processing is becoming a necessary technique to efficiently store, manage, and analyze massive data obtained by social media contents. NoSQL is one of databases that efficiently handle Big Data compared to the traditional database that has the limitation to manipulate Big Data. Log-structured key/value store, widely used in NoSQL, basically stores data into the disk storage in batch writing. Since this batch writing of the key/value store does not overwrite data in place, many data are accumulated in several places. Although it improves the write performance, the read performance decreases because the key/value store requires many accesses to widely-spread data. In order to address this problem, we propose T-tree index structure to reduce the search time by avoiding exploring contents stored in distributed many files. Finally, we show the performance improvement through the experimental results.


Electronics, Information and Communications (ICEIC), 2014 International Conference on | 2014

A B-tree employing a cascade memory node on NAND flash SSDs

Bo-kyeong Kim; Chang-Jun Yeo; Dong-Ho Lee

A NAND flash SSD has been widely used as a storage device because of its fast access speed and low power consumption. It inherits distinctive features such as no in-place update because it consists of NAND flash arrays. When implementing a B-tree on flash SSDs, the performance degradation occurs due to intensive updates in the same node. In this paper, we propose a B-tree employing a cascade memory node on NAND flash SSDs in order to reduce intensive overwrites. Our B-tree keeps the modified data so as to delay write operations and performs the flushing of memory node in batch process. Through the experiment, we show the superiority of our B-tree compared to related techniques.


Journal of KIISE:Databases | 2012

An Efficient B-Tree Using Lazy Update on Flash Memory

Bo-kyeong Kim; Min-Hee Yoo; Dong-Ho Lee

Collaboration


Dive into the Bo-kyeong Kim's collaboration.

Top Co-Authors

Avatar
Top Co-Authors

Avatar
Top Co-Authors

Avatar
Top Co-Authors

Avatar
Top Co-Authors

Avatar
Top Co-Authors

Avatar
Top Co-Authors

Avatar
Top Co-Authors

Avatar

Sang-Won Lee

Sungkyunkwan University

View shared research outputs
Researchain Logo
Decentralizing Knowledge