• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.11 no.12
• /
• pp.4997-5006
• /
• 2010

A flash-memory-based SSDs(Solid State Disks) are one of the best media to support portable and desktop computers' storage devices. Their features include non-volatility, low power consumption, and fast access time for read operations, which are sufficient to present flash memories as major database storage components for desktop and server computers. However, we need to improve traditional storage management schemes based on HDD(Hard Disk Drive) and RAID(Redundant array of independent disks) due to the relatively slow or freezing characteristics of write operations of SSDs, as compared to fast read operations. In order to achieve this goal, we propose a new storage management scheme called Hetero-Mirroring based on traditional HDD mirroring scheme. Hetero-Mirroring-based scheme improves RAID-1 operation performance by balancing write-workloads and delaying write operations to avoid SSD freezing. Our test results show that our scheme significantly reduces the write operation overheads and freezing overheads, and improves the performance of traditional SSD-RAID-1 scheme by 18 percent, and the response time of the scheme by 38 percent.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.7 no.1
• /
• pp.158-163
• /
• 2003

When the number of disks array systems that contain multiple disk drives, system performance is limited by a bottleneck at a centralized controller at a communication path than uses a bus. A redundant array of inexpensive disks(RAID) consists of many disks to enable high performance and large capacity. We evaluate a scalable architecture called Autonomous network, in which the controller functions are distributed to all disk drives and each disk has autonomy in processing its tasks. Disks drives enable better scalability and more effective utilization of system resources than with a hierarchical system. Autonomous network provided high read/write performance throughput in proportion to the number of disks.

• Proceedings of the IEEK Conference
• /
• 1998.10a
• /
• pp.495-498
• /
• 1998

Mobile computers have restrictions for size, weight, and power consumption that are different from traditional workstations. Storage device must be smaller, lighter. Low power consumed storage devices are needed. At the present time, flash memory device is a reasonable candidate for such device. But flash memory has drawbacks such as bulk erase operation and slow program time. This causes of worse average write performances. This paper suggests a storage method which improves write performance.

• Journal of the Korea Society of Computer and Information
• /
• v.15 no.2
• /
• pp.19-26
• /
• 2010

The Windows registry is a hierarchical database where the configuration data of a system or application programs is stored. In this paper, we presented and implemented a registry virtualization algorithm and measured its performance. The registry virtualization algorithm presented in the paper is called Copy-One-level On Write-Open(COOWO) that is a modified version of general Copy On Whte(COW) method to make it suitable for registry virtualization. In this paper, we implemented the proposed algorithm as a dynamically loadable library in Windows and applied it to many Windows application programs. This paper is meaningful since we described a registry virtualization algorithm in detail in situation where we can not find papers that describe the registry virtualization in detail, and we could find the performance of the algorithm can be used in the real applications.

• KIISE Transactions on Computing Practices
• /
• v.22 no.4
• /
• pp.163-169
• /
• 2016

Lustre is a scalable, distributed file system, which is popular in the field of high-performance computing. Recently, the advent of SSD has enabled high-performance storage hardware, but software development requires further improvement. In this paper, we analyzed performance of the Lustre system using SSD via extensive experimentation. We compared performance of Lustre on SSDs and HDDs in terms of file read/write throughputs and metadata access latencies. Our experimental results showed that 1) SSDs improve metadata access performance due to fast random read/write access of SSD characteristics, and 2) SSD are benefited to a greater extent under multiple threads and large numbers of small sized files.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2010.07a
• /
• pp.411-414
• /
• 2010

본 논문은 비디오 처리를 위한 고성능의 메모리 제어기를 설계하였다. 메모리 제어기는 arbiter에 의해 제어되며 이것은 메모리 억세스를 요구하는 모듈들의 요구 신호를 받아 데이터를 전송하는 역할을 해주게 된다. 구현된 메모리 제어기는 버스를 사용하기 위한 승인을 받기 위해서 마스터와 신호를 주고 받는 MAU블록, grant 신호를 디코딩하고 컨트롤 신호의 상태를 정의한 arbiter 블록, SDRAM의 ac parameter를 저장하고 bank의 준비 여부, read/write 가능 여부, precharge와 refresh의 가능 여부를 확인하여 system과 read/write가 준비되었다는 신호를 출력, SDRAM의 실질적인 입력신호를 생성하는 memory accelerator 블록, 생성된 입력신호를 저장하고 마스터에서 직접 write data를 입력 받는 memory I/F 블록으로 구성된다. 이 메모리 제어기는 174.28MHz의 주파수로 동작하였다. 본 설계는 VHDL을 이용하여 설계되었고, ALTERA의 Quartus II를 이용하여 합성하였다. 또한 ModelSim을 이용하여 설계된 회로를 검증하였다. 구현된 하드웨어는 StatixIII EP3SE80F1152C2 칩을 사용하였다.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.11 no.2
• /
• pp.117-123
• /
• 2016

In general, PCM (Phase Change Memory) is unsuitable as a main memory because it has limitations: high read/write latency and low endurance. However, the DRAM&PCM hybrid memory with the same level is one of the effective structures for a next generation main memory because it can utilize an advantage of both DRAM and PCM. Therefore, it needs an effective page management method for exploiting each memory characteristics dynamically and adaptively. So we aim reducing an access time and write count of PCM by using an effective page replacement. According to our simulation, the proposed algorithm for the DRAM&PCM hybrid can reduce the PCM access count by around 60% and the PCM write count by 42% given the same PCM size, compared with Clock-DWF algorithm.

• Transactions of the Society of Information Storage Systems
• /
• v.7 no.2
• /
• pp.47-52
• /
• 2011

This paper presents a method to improve vibration characteristics of servo track write (STW) fixture. STW fixtures supported by flexible mounts are subject to various vibration sources. Using Finite element analysis (FEA) vibration modes of the fixture are identified. The FEA results suggest certain vibration modes be reduced through design change of flexible mounts to improve vibration responses of the fixture. Based on layered flexible mounts theory a parametric study on shear and bending stiffness is performed to obtain a suitable flexible mount design leading to increased resistance to rocking motion. Experiments confirm improvement of vibration characteristics and drive performance through new mounts design.

• Transactions of the Society of Information Storage Systems
• /
• v.6 no.2
• /
• pp.41-46
• /
• 2010

Recently, Flash Memory Drives are one of the best media to support portable and desktop computers' storage devices. Their features include non-volatility, low power consumption, and fast access time for read operations, which are sufficient to present flash memories as major database storage components for desktop and server computers. However, we need to improve traditional storage management schemes based on HDD(Hard Disk Drive) and RAID(Redundant Array of Independent Disks) due to the relatively slow or freezing characteristics of write operations of SSDs, as compared to fast read operations. In order to achieve this goal, we propose a new storage management scheme called Hetero-Mirroring based on traditional HDD mirroring scheme. Hetero-Mirroring-based storage management improves RAID-1 operation performance by balancing write-workloads and delaying write operations to avoid SSD freezing.

• Journal of the Semiconductor & Display Technology
• /
• v.21 no.3
• /
• pp.1-6
• /
• 2022

Modern computer systems usually have special hardware for operations used in deep learning workload even edge computing environment. Non-volatile memories (NVMs) have been considered for alternative memory storage because they consume little static energy and occupy small area. However, there is a problem for NVMs to be directly adopted. An NVM cell has limited write endurance, so that the lifetime of NVM-based memory system is much shorter than that of conventional memory system. To overcome this problem for the deep learning system, this paper proposes a novel method to extend the lifetime based on the analysis of the deep learning workloads. If an incoming block has more than a predefined number of frequently used values, the cacheline is defined as write friendly block. During the victim selection, the cacheline has lower possibility to be chosen as victim. The experimental results show that the lifetime is increased by about 50% and energy consumption is decreased by 3% with a little performance hurt.