Solid State Drives Research Articles

Advanced Data Mining of SSD Quality Based on FP-Growth Data Analysis

Storage devices in the computer industry have gradually transformed from the hard disk drive (HDD) to the solid-state drive (SSD), of which the key component is error correction in not-and (NAND) flash memory. While NAND flash memory is under development, it is still limited by the “program and erase” cycle (PE cycle). Therefore, the improvement of quality and the formulation of customer service strategy are topics worthy of discussion at this stage. This study is based on computer company A as the research object and collects more than 8000 items of SSD error data of its customers, which are then calculated with data mining and frequent pattern growth (FP-Growth) of the association rule algorithm to identify the association rule of errors by setting the minimum support degree of 90 and the minimum trust degree of 10 as the threshold. According to the rules, three improvement strategies of production control are suggested: (1) use of the association rule to speed up the judgment of the SSD error condition by customer service personnel, (2) a quality strategy, and (3) a customer service strategy.

Field source extraction of an ESD generator and its application to system-level ESD analysis in a solid-state storage system

System-level electrostatic discharge (ESD) noise due to electromagnetic fields radiating from an ESD generator were measured and analyzed in a solid-state drive (SSD) storage system. The field sources of an ESD generator were efficiently modeled from the measured tangential magnetic fields in front of a ESD generator. The extracted field sources were incorporated into a full-wave solver, and the fields inside the disk-array enclosure (DAE) were simulated and validated by comparing to measured fields. The field source models of the ESD generator were then also utilized to predict transient ESD-induced noise voltages in a simplified SSD. Finally, the coupling mechanism and the effects of the DAE geometry were investigated.

Self-Adapting Channel Allocation for Multiple Tenants Sharing SSD Devices

Solid-state drives (SSDs) have been widely deployed in high-performance data center environments, where multiple tenants usually share the same hardware. However, traditional SSDs distribute the users&#x2019; incoming data uniformly across all SSD channels, which leads to numerous access conflicts. Meanwhile, SSDs that blindly allocate one or several channels to one tenant sacrifice device parallelism and capacity. When SSDs are shared by tenants with different access patterns, inappropriate channel allocation results in SSD performance degradation. In this article, we propose a self-adapting channel allocation mechanism, named SSDKeeper, for multiple tenants that share one SSD. SSDKeeper employs a machine learning-assisted algorithm to take full advantage of SSD parallelism while providing performance isolation. By collecting multitenant access patterns, SSDKeeper predicts an optimal channel allocation strategy for multiple tenants using the well-trained model. To further consume the blocks in different channels evenly, SSDKeeper equips with a novel channel swap scheme to prolong the SSD lifespan. Comparing with traditional SSDs, SSDKeeper reduces the overall latency of read and write by 12.6&#x0025; and the lifespan is prolonged up to <inline-formula> <tex-math notation="LaTeX">$3.7\times $ </tex-math></inline-formula>.

A 1.8-Gb/s/Pin 16-Tb NAND Flash Memory Multi-Chip Package With F-Chip for High-Performance and High-Capacity Storage

This article presents a 1.2-V, 1.8-Gb/s/pin 16-Tb NAND flash memory multi-chip package incorporating 16 dies of 1-Tb NAND flash memory and the third-generation F-chip. The proposed third-generation F-chip is developed to meet the performance requirements of a high-capacity storage device that adopts a PCIe Gen four-host interface for higher data throughput. It is implemented with dual bi-directional transceiver architecture and signal retiming scheme to maximize the valid data window opening on solid-state drive (SSD) channels. Also, it facilitates training between F-chip and NAND using an on-chip delay-locked loop whose locking is proposed in strobe-based NAND systems to achieve sufficient signal integrity (SI) of the in-package channel at a speed of 1.8 Gb/s/pin. Embedded built-in self-test evaluates un-selected paths and determines if re-training is required without losing data throughput performance. This work achieves a 35% improvement in the I/O operational speed performance and a 23% reduction in the I/O power consumption in comparison with the previous generations.

EMT: Elegantly Measured Tanner for Key-Value Store on SSD

With the emergence of big data era, NoSQL key-value database is considered as a promising candidate for replacing relational database management system (RDBMS). As cost per GB of flash memory is getting closer to HDD, high performance solid state drive (SSD) is regarded as the best substitute of HDD. However, there are still some issues that need to be taken into consideration. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">et al.</i> have reported that applying key-value store to SSD with conventional FTL would incur internal fragmentation and further cause the degradation of device lifespan. Although they proposed KVFTL to deal with the above issues, their work gave rise to the read amplification problem. In this article, we investigate the root cause of the read amplification problem and propose elegantly measured tanner FTL (EMT-FTL) to mitigate internal fragmentation and read amplification with acceptable memory overhead. Different from KVFTL that slices a variable-sized value into multiple partitions (up to 20) of different sizes and manages them as a linked chain, EMT-FTL slices a value into 16 KiB full partitions with its remaining bytes treated as a fragment partition and appended to the fragment buffer. For a 64 GiB SSD with 16 KiB pages, the overall memory usage of EMT-FTL is only 8.81% of KVFTL. The experiments showed that EMT-FTL achieved almost the optimal space utilization as KVFTL did under most of traces and averagely improved the <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">get</monospace> performance by 78.57% (compared with KVFTL) under the traces with request sizes ranging from 1 byte to 16 KiB.

HDFTL: An On-Demand Flash Translation Layer Algorithm for Hybrid Solid State Drives

NAND flash-based solid-state drives (SSDs) have been widely used in consumer electronic products such as smartphones, digital video recorders, and laptops. The flash translation layer (FTL) is crucial to the performance of SSDs. In this article, an on-demand FTL, briefly as HDFTL, is proposed for hybrid SSDs built with SLC+MLC flash memory. HDFTL divides the SLC into data blocks and translation blocks and uses all MLC as data blocks. All mapping items are stored in the translation block of the SLC, and partial mapping items are loaded into the cached mapping table of the FTL as needed. The contributions of HDFTL are as follows: (1) Unify the address mapping method of SLC and MLC areas to simplify the FTL design of hybrid SSDs; (2) Split the cached mapping table into a hot-write mapping table and a normal mapping table and use time locality to solve hot-write identification problem in hybrid SSDs; (3) According to the relative wear rate of SLC and MLC areas, adaptively adjust the size of the hot-write mapping table and the normal mapping table. Experimental results show that the performance of HDFTL is better than that of the existing FTLs for hybrid SSDs.

Layer-Aware Request Scheduling for 3D Flash-Based SSDs

Recent development on 3D flash memories largely promotes the wide application of Solid-State Drives (SSDs) by providing larger capacity from vertically-stacked layers. However, there exist speed variations across different layers because of manufacturing process variations or physical designs, which induces new challenges to fully explore the advantages of existing SSDs, e.g. the parallel structure. This paper investigates the effect of speed variation on the parallel performance of SSDs. To balance chips’ workloads, traditional method selects chips in a round-robin way. As chip queue time can be estimated by some main factors, the chip also can be chosen in a greedy way. However, because of the layer speed variation, queue time estimation model should be modified. This paper first establishes a new queue time estimation model with the awareness of the flash layer information. Then the model is used to estimate the chip queue time and to direct write requests into the chip with the least queue time. The key idea is to largely reduce queue time of each write, thus reducing the average SSD response time. Finally, this new request redirection method is evaluated on SSDsim with real world workloads. Experimental results show that our model can estimate the queue time more accurately and our new request redirection method can improve 8.3% of write queue time on average under the situation of 4 times speed variation among 16 layers.

Optimizing Key-Value Stores for Flash-Based SSDs via Key Reshaping

Key-Value store (KV store) is becoming widely popular in both academia and industry due to its fast performance and simplicity in data management. To improve the performance of KV stores, recent Serial Advanced Technology Attachment (SATA) and Non-Volatile Memory express (NVMe) Solid-State Drives (SSDs) have been widely adopted. In contrast to the existing Hard-Disk Drives (HDDs), SSDs have unique characteristics which must be carefully considered to exploit the full performance. For example, due to the erase before write constraint, the access pattern of workloads impacts the performance and endurance of SSDs. Thus, the performance of SSD with the sequential workload is higher than that with the random workload. In this paper, we propose a key reshaping technique to improve the performance of KV stores with high performance storage devices. By reshaping keys, our scheme allows KV stores to process the random insert requests into sequential insert requests, improving request processing and Input/Output (I/O) performance. Our experimental results show that the proposed scheme can improve the performance of KV store by up to 106% and 281% compared with the existing scheme, in the case of SATA and NVMe SSDs, respectively.

High-speed playback of spatiotemporal division multiplexing holographic 3D video stored in a solid-state drive using a digital micromirror device

We propose a high-speed playback method for the spatiotemporal division multiplexing electroholographic three-dimensional (3D) video stored in a solid-state drive (SSD) using a digital micromirror device. The spatiotemporal division multiplexing electroholography prevents deterioration in the reconstructed 3D video from a 3D object comprising many object points. In the proposed method, the stored data is remarkably reduced using the packing technique, and the computer-generated holograms are played back at high speed. Consequently, we successfully reconstructed a clear 3D video of a 3D object comprising approximately 1,100,000 points at 60 frames per second by reducing the reading time of the stored data from an SSD.

An Early-Life NAND Flash Endurance Prediction System

NAND flash memory – ubiquitous in today’s world of smart phones, SSDs (solid state drives), and cloud storage – has a number of well-known reliability problems. NAND data contains bit errors, which require the use of error correcting codes (ECCs). The raw bit error rate (RBER) increases with program-erase (P-E) cycling, and the number of P-E cycles the device can withstand before the RBER exceeds the ECC capability is called its endurance . ECC operates on data stored in a sector of NAND, and there is a large variation in the endurance of sectors within a device and across devices, resulting in excessively conservative endurance specifications. This research shows, for the first time, that a sector’s true endurance can be predicted with remarkable accuracy, using a combination of the sector’s location within the device, and measurements taken at the very beginning of life. Real-world data is gathered on millions of NAND sectors using a custom-built test platform. Optimised machine learning classification models are built from the raw data to predict if a sector will pass or fail to a fixed ECC threshold, after a target P-E cycling level has been reached. A novel technique is demonstrated that uses different ECC thresholds for model training and testing, which allows the models to be tuned so that they never misclassify samples that would fail. This eliminates ECC failures and data loss, allowing simpler, less expensive ECC schemes to be used for modern NAND devices. It also enables significant endurance extensions to be achieved.

Two-Stage In-Storage Processing and Scheduling for Pattern Matching Applications

In-storage processing technology allows applications to run on embedded processors and accelerators inside solid-state drives (SSDs) for efficient computing distribution. Especially, in pattern matching applications, in-storage computing can be processed quickly due to low data access latency, and the number of I/Os can be reduced by returning only a small amount of results to the host system after processing. Previously proposed in-storage processing is separated into three phases: command decoding, data access, and data processing. In this case, data processing is strictly isolated from data access, and the isolation constraints the utilization of storage. Merging data access and data processing among the phases can enhance the utilization of storage. To efficiently merge them, we propose two-stage in-storage processing and scheduling, especially for the pattern matching application. The first stage processing during data access reduces the second stage processing latency. Also, leveraging the pattern matching results of the first stage processing, our scheduler prioritizes key requests that should return the results to the host system so that they are completed earlier than non-key requests. The proposed scheduling reduces the response time of in-storage processing requests by 52.6 % on average.

USING CLOUD PLATFORMS TO BUILD DISTRIBUTED LEARNING MANAGEMENT SYSTEMS

Distributed systems have problems with downtime, data loss during malfunctions, scalability and efficient use of computing resources. At the same time in the learning and training process, the use of a distributed system has the advantage of data processing: storage of information about students, construction of training courses, verification of passed material, etc. The problems of scaling and efficient use of resources in distributed learning management systems are investigated in this research. Cloud platforms for hosting the system, such as Amazon Web Services, Microsoft Azure, Google Cloud Platform and DigitalOcean are reviewed. Problems and features of a scalability in cloud computing are discussed. Methods, scaling and load balancing algorithms for the efficient use of computing resources are proposed. According to the list of advantages, the DigitalOcean platform was selected for the investigation. DigitalOcean provides cloud servers that can be used for quick creation of the new virtual machines for the projects. These servers allow to fully control the web hosting environment at the same time that the user pays only for the resources used. The main goal of DigitalOcean is to use a solid-state drive (SSD) to create a user-friendly platform that will allow clients to migrate projects to and from the cloud, increasing productivity with high speed and efficiency. As a result of analyzing information on existing technologies, approaches and methods for using cloud platforms in distributed systems, they have been applied to develop a solution to reduce downtime for a distributed adaptive Learning Management System (LMS). It is concluded that the use of cloud platforms for the construction of distributed LMS a practice that allows to use only the required amount of computing capacity. It is proven, that the implementation of the proposed solution into the work of adaptive LMS will improve its efficiency by reducing the time of the content delivering.

An Experimental Analysis of the Use of Different Storage Technologies on a Relational DBMS

Traditional Database Management Systems (DBMSs) are built with the premise that magnetic disks such as hard disks drives (HDDs) store the data. Recently, several alternatives to HDDs have emerged, such as the solid-state drives (SSDs) based on non-volatile memory (NVM) technology such as 3D XPoint and the new generations of dynamic random access memories (DRAMs). Different characteristics of these storage technologies may impact the performance of DBMSs. In this work, we analyze the performance of a DBMS using three storage technologies as data locations:HDD, SSD NVM, and DRAM, as well as a hybrid way combining all three. To do this, we use two workloads, analytical and transactional, and we observe throughput as well as latency. After, we discuss the reasons for the results obtained for each type of storage. We also show that the query processing can benefit from the different characteristics of each storage technology to perform faster queries and, finally, we analyze the benefits of using a hybrid storage system.

An intelligent energy efficient storage system for cloud based big data applications

Storage technology has emerged as an indispensable paradigm for processing various applications in cloud data centers. The storage infrastructure consisting of Hard Disk Drives (HDDs) and Solid-State Drives (SSDs) accounts for high energy consumption. Also, the trade-offs between HDDs and SSDs in terms of cost and energy consumption are extremely high. Therefore, disk-based storage subsystems need to be more energy efficient. This paper proposes an intelligent energy-efficient hybrid disk storage system. The proposed system recognizes the frequently used data from traces of applications. Replica management along with data layout is used to allocate frequently used files in hot disks and the other files in cold disks. The request is executed using an intelligent scheduling technique that searches and selects the disk based on its states. The scheduling technique selects either an idle or active disk without spinning the standby disks. The selection procedure also considers minimum waiting time for the active disk and maximum remaining idle time for the idle disk. The proposed system has been implemented by adding disk management in the cloud environment, which proved to be highly effective in achieving a 39% power savings with an 18.26% decrease in execution time.

SACRO : Solid state drive‐assisted chunk caching for restore optimization

AbstractBetter duplicate elimination performance causes higher fragmentation which leads to degraded restore performance. As a result, restore performance needs to be optimized either through strengthening locality by selective rewriting and/or making the best use of the limited available memory through cache optimization. In this paper, we explore SACRO, SSD Assitsted Chunk Caching for Restore Optimization. It avoids the need to repetitively access chunk containers in disk by using SSD (Solid State Drives) as a secondary chunk cache. An FRT (Future Reference Table) is constructed from the recipe of a backup stream and a Future Reference Count entry in the FRT is utilized to assign priorities to chunks as they are accessed during restoration. These priority values coupled with access distances are used to decide to cache a chunk either in memory or the SSD. The restore performance of SACRO is shown to significantly outperform other restoring approaches which utilize chunk caching. For one dataset, at 4 MB cache size and 4 MB Look Ahead Window size its restore factor is 3.5 MB/container_access while ALACC and LRU record 3.4 MB/container_access and 0.9 MB/container_access respectively. Moreover, SACRO can be integrated to already existing deduplication systems with very limited amount of modification done on the deduplication system.

Data Bucket-Based Fragment Management for Solid State Drive Storage System

The current storage mechanism considered little in data’s keeping characteristics. These can produce fragments of various sizes among data sets. In some cases, these fragments may be serious and harm system performance. In this paper, we manage to modify the current storage mechanism. We introduce an extra storage unit called data bucket into the classical data manage architecture. Next, we modify the data manage mechanism to improve our designs. By keeping data according to their visited information, both the number of fragments and the fragment size are greatly reduced. Considering different data features and storage device conditions, we also improve the solid state drive (SSD) lifetime by keeping data into different spaces. Experiments show that our designs have a positive influence on the SSD storage density and actual service time.

Building Reliable Massive Capacity SSDs through a Flash Aware RAID-Like Protection

The demand for mass storage devices has become an inevitable consequence of the explosive increase in data volume. The three-dimensional (3D) vertical NAND (V-NAND) and quad-level cell (QLC) technologies rapidly accelerate the capacity increase of flash memory based storage system, such as SSDs (Solid State Drives). Massive capacity SSDs adopt dozens or hundreds of flash memory chips in order to implement large capacity storage. However, employing such a large number of flash chips increases the error rate in SSDs. A RAID-like technique inside an SSD has been used in a variety of commercial products, along with various studies, in order to protect user data. With the advent of new types of massive storage devices, studies on the design of RAID-like protection techniques for such huge capacity SSDs are important and essential. In this paper, we propose a massive SSD-Aware Parity Logging (mSAPL) scheme that protects against n-failures at the same time in a stripe, where n is protection strength that is specified by the user. The proposed technique allows for us to choose the strength of protection for user data. We implemented mSAPL on a trace-based simulator and evaluated it with real-world I/O workload traces. In addition, we quantitatively analyze the error rates of a flash based SSD for different RAID-like configurations with analytic models. We show that mSAPL outperforms the state-of-the-art RAID-like technique in the performance and reliability.

Detecting firmware modification on solid state drives via current draw analysis

Solid State Drives (SSDs) have gained significant market share among data storage options in recent years due to increased speed and durability. But when compared to Hard Disk Drives (HDDs), SSDs contain additional complexity which must be managed in firmware. Some manufacturers make firmware updates available, but their proprietary protections leave end users unable to verify the authenticity of the firmware post installation. This means that attackers who are able to get a malicious firmware version installed on a victim SSD are able to operate with impunity, as the owner will have no tools for detection. We use a method for performing side channel analysis of the current drawn by an SSD to compare its behavior while running genuine firmware against its behavior when running modified firmware. We further test this method for robustness against changes in external factors such as temperature and supplied power. In each case, we train a binary classifier with samples of genuine as well as modified firmware activity and are able to discriminate between them with over 90% accuracy in most experiments. Solid State Drives are trusted to store and protect critical data, so verification of SSD firmware is an important step towards having trust and confidence in the growing landscape of embedded devices used for critical operations.

Improving SSD Read Latency via Coding

We study the potential enhancement of the read access speed in high-performance solid-state drives (SSDs) by coding, given speed variations across the multiple flash interfaces and assuming occasional local memory failures. Our analysis is based on a queuing model that incorporates both read request failures and NAND element failures. The NAND element failure in the present context reflects various limitations on the memory element level such as bad blocks, dies or chips that cannot be corrected by error control coding (ECC) typically employed to protect pages read off the NAND cells. Our analysis provides a clear picture of the storage-overhead and read-latency trade-offs given read failures and NAND element failures. We investigate two different ways to mitigate the effect of NAND element failures using the notion of multi-class jobs with different priorities. A strong motivation for this work is to understand the reliability requirement of NAND chip components given an additional layer of failure protection, under the latency/storage-overhead constraints.

Wear-Aware Out-of-Order Dynamic Scheduling for NAND Flash-Based Consumer Electronics

Because of the excellent performance and decreasing price of NAND flash, NAND flash-based solid-state drive (SSD) has been widely used as a storage system in consumer electronics. Moreover, modern consumer SSDs usually adopt a multichannel parallel structure. Thus, how to fully utilize the internal parallelism through IO scheduling is a key problem. The existing IO schedulers fail to solve the following: 1) the utilization of flash translation layer information to optimize the internal parallelism is not sufficient; 2) the wear of flash memory is not considered when allocating write requests. This article proposes a Wear-aware Out-of-order Dynamic Scheduling Algorithm (WODSA). First, according to the information of the flash translation layer, the max-parallel-based scheduling strategy is proposed to schedule the read requests to maximize read parallelism. Second, the wear-aware dynamic write allocation strategy is proposed based on the idle/busy state and wear degree of channels and chips. WODSA allocates write requests to channels and chips with less wear preferentially in a maximized parallel manner to achieve write parallelism and active dynamic wear-leveling. Experimental results show that compared with existing IO schedulers and dynamic wear-leveling algorithms, WODSA can improve both the average response time and wear-leveling in block-level and channel-level.