Computer data storage refers to the digital preservation and backup/recovery of files, documents, and other forms of information stored digitally for safe keeping and recovery in case any of this information becomes lost or deleted.
Most computers utilize a storage hierarchy that prioritizes fast volatile memory technologies closer to the central processing unit while slower non-volatile persistent technologies are used as secondary and off-line storage, such as USB drives, floppy disks or magnetic tape.
Hard Disk Drives (HDDs)
Hard disk drives (HDDs) are mechanical drives used to store and retrieve digital data. As part of computer systems, HDDs serve as the primary storage devices in desktops and laptops; their non-volatile memory helps preserve stored information even when power is turned off, and programs instruct their operating systems to access this drive when reading or writing data is necessary.
An HDD’s physical components, known as platters, take the shape of discs. Constructed of rigid materials – typically aluminum or glass – covered with magnetic layers to store data, these platters spin around a central hub at 5,400-15,000 RPM while being controlled by CPU and motherboard to move across them with actuator arms carrying read/write heads which act on every individual platter by moving across their electrical charge charges or by the actuator arm’s position over areas within each platter’s surface – eventually translating this data into binary information via its read/write heads.
Engineers initially designed HDDs using longitudinal recording, in which sectors were aligned horizontally on the platter. But each sector became problematic as its magnetic charge fluctuated based on temperature changes, leading to data corruption. To address this issue, perpendicular recording was introduced: this method stacks sectors vertically for three times more storage capacity than longitudinal recording.
Modern HDDs stand out from their predecessors by being capable of storing massive amounts of data in a relatively small surface area, which allows for high storage capacities at reasonable costs. A standard HDD can store up to 2 TB worth of information – although this figure may seem significantly smaller when operating systems, file structures and redundancy procedures take their share from each drive.
An internal hard drive needs a data connector in order to communicate with other parts of the computer, whether this means enhanced integrated drive electronics (IDE), Serial ATA (SATA) or even newer interfaces such as SAS, SCSI or NVMe.
Solid-State Drives (SSDs)
An SSD stores data in semiconductor blocks that hold one or more bits of information. It is significantly faster than its HDD counterpart and requires significantly less power; its lack of moving parts also makes it more resistant to physical shock, quieter, and energy efficient than an HDD – though SSDs tend to cost more.
An SSD comprises two main components: its flash controller and NAND flash memory chips, both designed to deliver superior read/write performance for sequential and random data requests. Furthermore, its circuit board contains empty spaces which can be filled by adding more NAND chips as its capacity grows; thus enabling its use across varying capacities levels with one single circuit board design.
SSDs connect to host systems through various interfaces depending on their storage technology; such as Serial ATA (SATA) or Serial Attached SCSI (SAS), or special-purpose interfaces like Fibre Channel and iSCSI. New drives use multi-level cell flash technology which increases lifespan of storage cells.
Many SSDs come equipped with built-in capacitors designed to save at least the FTL mapping table in case of unexpected power outage, while enterprise class SSDs like Intel DC S3700 series feature batteries or supercapacitors for even greater reliability.
An SSD can typically store up to 2TB of data. Models designed specifically for consumer devices like smartphones and tablets as well as larger capacities for server applications are all available. Incorporating wear leveling technology that reduces how often memory cells are programmed/erased helps extend its lifespan further.
An SSD measures its storage capacity in gigabytes or terabytes, and increasing that capacity can be done using a log-structured file system with support for TRIM, which recycles memory blocks that have been cleared out by using log file system TRIM command.
As they don’t use mechanical spinning disks, SSDs are much faster at booting up and loading applications than HDDs. Furthermore, they’re generally quieter, consume less power, and come in smaller forms factors than their HDD counterparts – often used in laptops, digital cameras and music players, smartphones, and tablet computers.
Flash Storage Memory Devices
Flash memory devices come in all shapes and sizes depending on their intended use case. A thumb drive designed to store media files is usually smaller and more compact than an SSD that’s intended as the main boot drive of a computer. They can even be customized with logos, text and images that promote businesses or brands; in addition to being faster and more durable than traditional hard drives.
Flash memory chips are non-volatile forms of storage that preserve information even when power is off, offering significant improvements over traditional magnetic disk storage media, which loses information upon being powered off. As such, flash memory has increasingly replaced magnetic storage media across a range of computer products from desktops and laptops to tablet PCs and cell phones.
There are two primary kinds of flash chips: NOR and NAND. While NOR is best-suited for code storage and execution, NAND excels at high capacity data storage. Both utilize floating-gate transistors – essential building blocks of digital electronics that regulate electricity flow – in each cell of their flash chip to control how electricity moves throughout them. Each flash cell contains both floating gates and control gates separated by an oxide layer; one stores electrons while the other changes their value from 0-1 depending on which way it’s pushed through its respective flash chip cells.
To read a flash chip, the source and drain of each transistor are connected via a bit line. NAND memory architecture allows more cells to be connected at one time, which reduces the number of ground wires and bit lines necessary to reach each cell and improves performance over NOR Flash memory which requires two separate grounds and bit lines for every two cells.
Both NOR and NAND flash storage can be electrically programmed and erased, making them more flexible than EEPROMs or older forms of memory. However, flash memory requires longer to erase than to program/write; its bit error rate is slightly higher than HDDs but its low power consumption and light weight make it popular choice for portable computer accessories.
Network-Attached Storage (NAS)
NAS (network attached storage) is a centralized file server where users can store and share files over WiFi or Ethernet networks. Also referred to as a “NAS box,” “NAS unit,” or NAS server, these devices resemble regular computers in that they contain multiple hard drives with various traditional magnetic and solid state technologies that serve as mass storage for users. In addition, their CPU manages these functions and offers additional redundancy enhancement features like mirroring or RAID implementations for added protection of user files.
Direct Attached Storage (DAS), in contrast, requires physically disconnecting and connecting it back to another computer before accessing its data. With NAS however, multiple computers can access one shared drive using its file system and server capabilities simultaneously – unlike DAS which must be physically removed before connecting back. Furthermore, depending on its configuration a NAS may support block level protocols like Fibre Channel, iSCSI and ATA over Ethernet in addition to file level protocols like CIFS/NFS/FTP for easy file sharing between computers.
There are three major types of network attached storage (NAS). Each meets specific user needs: from home and small business users seeking local shared storage for just a few client systems up to several terabytes; midmarket NAS caters more closely to SMBs and enterprise workgroups and can include advanced features to address security and regulatory compliance requirements; finally, high end NAS are designed specifically to support large enterprises needing secure workgroup storage environments that address security and regulatory compliance compliance requirements.
At the top end of the market, enterprise NAS meets the rigorous needs of large organizations by supporting petabytes of storage and thousands of clients with rapid access, clustering capabilities and performance optimization features.
NAS, DAS and SAN are three primary storage architectures used by users for storage needs. When seeking cost-effective, highly scalable solutions that can be managed by IT teams easily, NAS may be ideal. However, companies with uncertain long-term storage needs might prefer cloud services or SAN storage solutions more; such as when processing large unstructured data sets or employing AI. In such situations, more robust systems with scale-out capabilities and built-in big data functionality might be necessary.
