A solid-state drive (SSD) is a type of solid-state storage device that uses integrated circuits to store data persistently. It is sometimes called semiconductor storage device, solid-state device, and solid-state disk.[1][2]
SSDs rely on non-volatile memory, typically NAND flash, to store data in memory cells. The performance and endurance of SSDs vary depending on the number of bits stored per cell, ranging from high-performing single-level cells (SLC) to more affordable but slower quad-level cells (QLC). In addition to flash-based SSDs, other technologies such as 3D XPoint offer faster speeds and higher endurance through different data storage mechanisms.
Unlike traditional hard disk drives (HDDs), SSDs have no moving parts, allowing them to deliver faster data access speeds, reduced latency, increased resistance to physical shock, lower power consumption, and silent operation.
Often interfaced to a system in the same way as HDDs, SSDs are used in a variety of devices, including personal computers, enterprise servers, and mobile devices. However, SSDs are generally more expensive on a per-gigabyte basis and have a finite number of write cycles, which can lead to data loss over time. Despite these limitations, SSDs are increasingly replacing HDDs, especially in performance-critical applications and as primary storage in many consumer devices.
SSDs come in various form factors and interface types, including SATA, PCIe, and NVMe, each offering different levels of performance. Hybrid storage solutions, such as solid-state hybrid drives (SSHDs), combine SSD and HDD technologies to offer improved performance at a lower cost than pure SSDs.
An SSD stores data in semiconductor cells, with its properties varying according to the number of bits stored in each cell (between 1 and 4). Single-level cells (SLC) store one bit of data per cell and provide higher performance and endurance. In contrast, multi-level cells (MLC), triple-level cells (TLC), and quad-level cells (QLC) store more data per cell but have lower performance and endurance. SSDs using 3D XPoint technology, such as Intel’s Optane, store data by changing electrical resistance instead of storing electrical charges in cells, which can provide faster speeds and longer data persistence compared to conventional flash memory.[3] SSDs based on NAND flash slowly leak charge when not powered, while heavily-used consumer drives may start losing data typically after one to two year in storage.[4] SSDs have a limited lifetime number of writes, and also slow down as they reach their full storage capacity.
SSDs also have internal parallelism that allows them to manage multiple operations simultaneously, which enhances their performance.[5]
Unlike HDDs and similar electromechanical magnetic storage, SSDs do not have moving mechanical parts, which provides advantages such as resistance to physical shock, quieter operation, and faster access times. Their lower latency results in higher input/output rates (IOPS) than HDDs.[6]
Some SSDs are combined with traditional hard drives in hybrid configurations, such as Intel's Hystor and Apple's Fusion Drive. These drives use both flash memory and spinning magnetic disks in order to improve the performance of frequently-accessed data.[7][8]
Traditional interfaces (e.g. SATA and SAS) and standard HDD form factors allow such SSDs to be used as drop-in replacements for HDDs in computers and other devices. Newer form factors such as mSATA, M.2, U.2, NF1/M.3/NGSFF,[9][10] XFM Express (Crossover Flash Memory, form factor XT2)[11] and EDSFF[12][13] and higher speed interfaces such as NVM Express (NVMe) over PCI Express (PCIe) can further increase performance over HDD performance.[3]
Traditional HDD benchmarks tend to focus on the performance characteristics such as rotational latency and seek time. As SSDs do not need to spin or seek to locate data, they are vastly superior to HDDs in such tests. However, SSDs have challenges with mixed reads and writes, and their performance may degrade over time. Therefore, SSD testing typically looks at when the full drive is first used, as the new and empty drive may have much better write performance than it would show after only weeks of use.[14]
The reliability of both HDDs and SSDs varies greatly among models.[15] Some field failure rates indicate that SSDs are significantly more reliable than HDDs.[16][17] However, SSDs are sensitive to sudden power interruption, sometimes resulting in aborted writes or even cases of the complete loss of the drive.[18]
Most of the advantages of solid-state drives over traditional hard drives are due to their ability to access data completely electronically instead of electromechanically, resulting in superior transfer speeds and mechanical ruggedness.[19] On the other hand, hard disk drives offer significantly higher capacity for their price.[6][20]
In traditional HDDs, a rewritten file will generally occupy the same location on the disk surface as the original file, whereas in SSDs the new copy will often be written to different NAND cells for the purpose of wear leveling. The wear-leveling algorithms are complex and difficult to test exhaustively. As a result, one major cause of data loss in SSDs is firmware bugs.[21][22]
While both memory cards and most SSDs use flash memory, they have very different characteristics, including power consumption, performance, size, and reliability.[39] Originally, solid state drives were shaped and mounted in the computer like hard drives.[39] In contrast, memory cards (such as Secure Digital (SD), CompactFlash (CF), and many others) were originally designed for digital cameras and later found their way into cell phones, gaming devices, GPS units, etc. Most memory cards are physically smaller than SSDs, and designed to be inserted and removed repeatedly.[39]
SSDs have different failure modes from traditional magnetic hard drives. Because solid-state drives contain no moving parts, they are generally not subject to mechanical failures. However, other types of failures can occur. For example, incomplete or failed writes due to sudden power loss may be more problematic than with HDDs, and the failure of a single chip may result in the loss of all data stored on it. Nonetheless, studies indicate that SSDs are generally reliable, often exceed their manufacturer-stated lifespan[40][41] and having lower failure rates than HDDs.[40] However, studies also note that SSDs experience higher rates of uncorrectable errors, which can lead to data loss, compared to HDDs.[42]
The endurance of an SSD is typically listed on its datasheet in one of two forms:
For example, a Samsung 970 EVO NVMe M.2 SSD (2018) with 1 TB of capacity has an endurance rating of 600 TBW.[44]
Recovering data from SSDs presents challenges due to the non-linear and complex nature of data storage in solid-state drives. The internal operations of SSDs vary by manufacturer, with commands (e.g. TRIM and the ATA Secure Erase) and programs like (e.g. hdparm) being able to erase and modify the bits of a deleted file.
The JEDEC Solid State Technology Association (JEDEC) has established standards for SSD reliability metrics, which include:[45]
In a distributed computing environment, SSDs can be used as a distributed cache layer that temporarily absorbs the large volume of user requests to slower HDD-based backend storage systems. This layer provides much higher bandwidth and lower latency than the storage system would, and can be managed in a number of forms, such as a distributed key-value database and a distributed file system. On supercomputers, this layer is typically referred to as burst buffer.
Flash-based solid-state drives can be used to create network appliances from general-purpose personal computer hardware. A write protected flash drive containing the operating system and application software can substitute for larger, less reliable disk drives or CD-ROMs. Appliances built this way can provide an inexpensive alternative to expensive router and firewall hardware.[citation needed]
SSDs based on an SD card with a live SD operating system are easily write-locked. Combined with a cloud computing environment or other writable medium, an OS booted from a write-locked SD card is reliable, persistent and impervious to permanent corruption.
In 2011, Intel introduced a caching mechanism for their Z68 chipset (and mobile derivatives) called Smart Response Technology, which allows a SATA SSD to be used as a cache (configurable as write-through or write-back) for a conventional, magnetic hard disk drive.[46] A similar technology is available on HighPoint's RocketHybrid PCIe card.[47]
Solid-state hybrid drives (SSHDs) are based on the same principle, but integrate some amount of flash memory on board of a conventional drive instead of using a separate SSD. The flash layer in these drives can be accessed independently from the magnetic storage by the host using ATA-8 commands, allowing the operating system to manage it. For example, Microsoft's ReadyDrive technology explicitly stores portions of the hibernation file in the cache of these drives when the system hibernates, making the subsequent resume faster.[48]
Dual-drive hybrid systems are combining the usage of separate SSD and HDD devices installed in the same computer, with overall performance optimization managed by the computer user, or by the computer's operating system software. Examples of this type of system are bcache and dm-cache on Linux,[49] and Apple's Fusion Drive.
The primary components of an SSD are the controller and the memory used to store data. Traditionally, early SSDs used volatile DRAM for storage, but since 2009, most SSDs utilize non-volatile NAND flash memory, which retains data even when powered off.[50][3] Flash memory SSDs store data in metal–oxide–semiconductor (MOS) integrated circuit chips, using non-volatile floating-gate memory cells.[51]
Every SSD includes a controller, which manages the data flow between the NAND memory and the host computer. The controller is an embedded processor that runs firmware to optimize performance, managing data, and ensuring data integrity.[52][53]
Some of the primary functions performed by the controller are:
The overall performance of an SSD can scale with the number of parallel NAND chips and the efficiency of the controller. For example, controllers that enable parallel processing of NAND flash chips can improve bandwidth and reduce latency.[55]
Micron and Intel pioneered faster SSDs by implementing techniques such as data striping and interleaving to enhance read/write speeds.[56] More recently, SandForce introduced controllers that incorporate data compression to reduce the amount of data written to the flash memory, potentially increasing both performance and endurance.[57]
Wear leveling is a technique used in SSDs to ensure that write and erase operations are distributed evenly across all blocks of the flash memory. Without this, specific blocks could wear out prematurely due to repeated use, reducing the overall lifespan of the SSD. The process moves data that is infrequently changed (cold data) from heavily used blocks, so that data that changes more frequently (hot data) can be written to those blocks. This helps distribute wear more evenly across the entire SSD. However, this process introduces additional writes, known as write amplification, which must be managed to balance performance and durability.[58][59]
Most SSDs use non-volatile NAND flash memory for data storage, primarily due to its cost-effectiveness and ability to retain data without a constant power supply. NAND flash-based SSDs store data in semiconductor cells, with the specific architecture influencing performance, endurance, and cost.[61]
There are various types of NAND flash memory, categorized by the number of bits stored in each cell:
Over time, SSD controllers have improved the efficiency of NAND flash, incorporating techniques such as interleaving, advanced error correction, and wear leveling to optimize performance and extend the lifespan of the drive.[63][64][65][66][67] Lower-end SSDs often use QLC or TLC memory, while higher-end drives for enterprise or performance-critical applications may use MLC or SLC.[68]
In addition to the flat (planar) NAND structure, many SSDs now use 3D NAND (or V-NAND), where memory cells are stacked vertically, increasing storage density while improving performance and reducing costs.[69]
Some SSDs use volatile DRAM instead of NAND flash, offering very high-speed data access but requiring a constant power supply to retain data. DRAM-based SSDs are typically used in specialized applications where performance is prioritized over cost or non-volatility. Many SSDs, such as NVDIMM devices, are equipped with backup power sources such as internal batteries or external AC/DC adapters. These power sources ensure data is transferred to a backup system (usually NAND flash or another storage medium) in the event of power loss, preventing data corruption or loss.[70][71] Similarly, ULLtraDIMM devices use components designed for DIMM modules, but only use flash memory, similar to a DRAM SSD.[72]
DRAM-based SSDs are often used for tasks where data must be accessed at high speeds with low latency, such as in high-performance computing or certain server environments.[73]
3D XPoint is a type of non-volatile memory technology developed by Intel and Micron, announced in 2015.[74] It operates by changing the electrical resistance of materials in its cells, offering much faster access times than NAND flash. 3D XPoint-based SSDs, such as Intel’s Optane drives, provide lower latency and higher endurance than NAND-based drives, although they are more expensive per gigabyte.[75][76]
Drives known as hybrid drives or solid-state hybrid drives (SSHDs) use a hybrid of spinning disks and flash memory.[77][78] Some SSDs use magnetoresistive random-access memory (MRAM) for storing data.[79][80]
Many flash-based SSDs include a small amount of volatile DRAM as a cache, similar to the buffers in hard disk drives. This cache can temporarily hold data while it is being written to the flash memory, and it also stores metadata such as the mapping of logical blocks to physical locations on the SSD.[81]
Some SSD controllers, like those from SandForce, achieve high performance without using an external DRAM cache. These designs rely on other mechanisms, such as on-chip SRAM, to manage data and minimize power consumption.[82]
Additionally, some SSDs use an SLC cache mechanism to temporarily store data in single-level cell (SLC) mode, even on multi-level cell (MLC) or triple-level cell (TLC) SSDs. This improves write performance by allowing data to be written to faster SLC storage before being moved to slower, higher-capacity MLC or TLC storage.[83]
On NVMe SSDs, Host Memory Buffer (HMB) technology allows the SSD to use a portion of the system’s DRAM instead of relying on a built-in DRAM cache, reducing costs while maintaining a high level of performance.[82]
In certain high-end consumer and enterprise SSDs, larger amounts of DRAM are included to cache both file table mappings and written data, reducing write amplification and enhances overall performance.[84]
Higher-performing SSDs may include a capacitor or battery, which helps preserve data integrity in the event of an unexpected power loss. The capacitor or battery provides enough power to allow the data in the cache to be written to the non-volatile memory, ensuring no data is lost.[85][86]
In some SSDs that use multi-level cell (MLC) flash memory, a potential issue known as "lower page corruption" can occur if power is lost while programming an upper page. This can result in previously written data becoming corrupted. To address this, some high-end SSDs incorporate supercapacitors to ensure all data can be safely written during a sudden power loss.[87]
Some consumer SSDs have built-in capacitors to save critical data such as the Flash Translation Layer (FTL) mapping table. Examples include the Crucial M500 and Intel 320 series.