• The Journal of the Korea Contents Association
• /
• v.17 no.7
• /
• pp.93-100
• /
• 2017

Recently, demands for high-performance flash-based storage devices (i.e., flash SSD) have rapidly grown in social network services, cloud computing, super-computing, and enterprise storage systems. The industry and academic communities made the NVMe specification for high-performance storage devices, and NVMe-based flash SSDs can be now obtained in the market. In this article, we evaluate performance of NoSQL databases that social network services and cloud computing services heavily adopt by using NVMe-based flash SSDs. To this end, we use NVMe SSD that Samsung Electronics recently developed, and the SSD used in this study has performance up to 3.5GB/s for sequential read/write operations. We use WiredTiger for NoSQL databases, and it is a default storage engine for MongoDB. Our experimental results show that log processing in NoSQL databases is a major overhead when high-performance NVMe-based flash SSDs are used. Furthermore, we optimize components of log processing and optimized WiredTiger show up to 15 times better performance than original WiredTiger.

• Journal of KIISE
• /
• v.42 no.12
• /
• pp.1480-1485
• /
• 2015

As SSDs are increasingly being used as high-performance storage or caches, attention is increasingly paid to the provision of SSDs with Quality-of-Service for I/O request streams of various applications in server systems. Since most SSDs are using the AHCI controller interface on a SATA bus, it is not possible to provide a differentiated service by distinguishing each I/O stream from others within the SSD. However, since a new SSD interface, the NVME controller interface on a PCI Express bus, has been proposed, it is now possible to recognize each I/O stream and schedule I/O requests within the SSD for differentiated services. This paper proposes Flash Operation Group Scheduling within NVME-based flash storage devices, and demonstrates through QEMU-based simulation that we can achieve a proportional bandwidth share for each I/O stream.

• KIISE Transactions on Computing Practices
• /
• v.24 no.3
• /
• pp.129-137
• /
• 2018

NVMe(Non-Volatile Memory Express) SSD(Solid State Drive) is a high-performance storage that makes use of flash memory as a storage cell, PCIe as an interface and NVMe as a protocol on the interface. It supports multiple I/O queues which makes it feasible to process parallel-I/Os on multi-core environments and to provide higher bandwidth than SATA SSDs. Hence, NVMe SSD is considered as a next generation-storage for data-center and cloud computing system. However, in the virtualization system, the performance of NVMe SSD is not fully utilized due to the bottleneck of the software I/O stack. Especially, when it uses I/O stack of the hypervisor or the host operating system like Xen and KVM, I/O performance degrades seriously due to doubled-I/O stack between host and virtual machine. In this paper, we propose a new I/O engine, called Direct-AIO (Direct-Asynchronous I/O) engine, that can access NVMe SSD directly for I/O performance improvements on QEMU emulator. We develop our proposed I/O engine and analyze I/O performance differences between the existed I/O engine and Direct-AIO engine.

• The Journal of the Korea Contents Association
• /
• v.18 no.4
• /
• pp.178-185
• /
• 2018

Recently, demands for computer systems with multicore CPUs and high-performance flash-based storage devices (i.e., flash SSD) have rapidly grown in cloud computing, surer-computing, and enterprise storage/database systems. Journaling file systems running on high-performance systems do not exploit the full I/O bandwidth of high-performance SSDs. In this article, we evaluate and analyze the performance of the Linux EXT4 file system with high-performance SSDs and multicore CPUs. The system used in this study has 72 cores and Intel NVMe SSD, and the flash SSD has performance up to 2800/1900 MB/s for sequential read/write operations. Our experimental results show that checkpointing in the EXT4 file system is a major overhead. Furthermore, we optimize the checkpointing procedure and our optimized EXT4 file system shows up to 92% better performance than the original EXT4 file system.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.12 no.3
• /
• pp.139-147
• /
• 2017

In recent years, non-volatile memory express (NVMe), a new host controller interface standard, has been adapted to overcome performance bottlenecks caused by the acceleration of solid state drives (SSD). Recently, performance breakthrough cases over AHCI based SATA SSDs by adapting NVMe based PCI Express (PCIe) SSD to servers and PCs have been reported. Furthermore, replacing legacy eMMC-flash storage with NVMe based storage is also considered for next generation of mobile devices such as smartphones. The Linux kernel includes drivers for NVMe support, and as the kernel version increases, the implementation of the NVMe driver code has changed. However, mobile devices are often equipped with older versions of Android operating systems (OSes), where the newest features of NVMe drivers are not available. Therefore, different features of different NVMe driver implementations are not well evaluated on Android OSes. In this paper, we analyze the response time of the NVMe driver for various Linux kernel version.

• Journal of KIISE
• /
• v.44 no.5
• /
• pp.469-475
• /
• 2017

The emerging NVMe-based multi-queue SSDs provides a high bandwidth by parallel I/O, i.e., each core performs I/O through its dedicated queue in parallel with other cores. To provide a bandwidth share for each application with I/O, a fair-share scheduler that provides a bandwidth share to each core is required. In this study, we proposed a multi-core scalable fair-queuing algorithm for multi-queue SSDs. The algorithm adopts randomization to minimize the inter-core synchronization overheads and provides a weight-proportional bandwidth share to each core. The results of our experiments indicated that the proposed algorithm gives accurate bandwidth partitioning and outperforms the existing FlashFQ scheduler, regardless of the number of cores for a Linux kernel with block-mq.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.14 no.2
• /
• pp.63-69
• /
• 2019

Recently, ultra-low latency flash storage devices such as Z-SSD and Optane SSD were introduced with the significant technological improvement in the storage devices which provide much faster response time than today's other NVMe SSDs. With such ultra-low latency, $10{\mu}s$, storage devices the cost of context switch could be an overhead during interrupt-driven I/O completion process. As an interrupt-driven I/O completion process could bring an interrupt handling overhead, polling or hybrid-polling for the I/O completion is known to perform better. In this paper, we analyze tail latency problem in a polling process caused by process scheduling in data center environment where multiple applications run simultaneously under one system and we introduce our adaptive polling selection technique which dynamically selects efficient processing method between two techniques according to the system's conditions.

• Journal of Internet of Things and Convergence
• /
• v.10 no.3
• /
• pp.19-24
• /
• 2024

Recent advances in large-scale data processing technologies such as big data, cloud computing, and artificial intelligence have increased the demand for high-performance storage devices in data centers and enterprise environments. In particular, the fast data response speed of storage devices is a key factor that determines the overall system performance. Solid state drives (SSDs) based on the Non-Volatile Memory Express (NVMe) interface are gaining traction, but new bottlenecks are emerging in the process of handling large data input and output requests from multiple hosts simultaneously. SSDs typically process host requests by sequentially stacking them in an internal queue. When long transfer length requests are processed first, shorter requests wait longer, increasing the average response time. To solve this problem, data transfer timeout and data partitioning methods have been proposed, but they do not provide a fundamental solution. In this paper, we propose a dual queue based scheduling scheme (DQBS), which manages the data transfer order based on the request order in one queue and the transfer length in the other queue. Then, the request time and transmission length are comprehensively considered to determine the efficient data transmission order. This enables the balanced processing of long and short requests, thus reducing the overall average response time. The simulation results show that the proposed method outperforms the existing sequential processing method. This study presents a scheduling technique that maximizes data transfer efficiency in a high-performance SSD environment, which is expected to contribute to the development of next-generation high-performance storage systems