Memory is a hardware device used to store data and programs in a computer system and is an important core of information processing. Its speed and capacity directly affect the overall computing performance.
Classification
Random Access Memory (RAM, Random Access Memory):It can be read and written, and disappears when power is turned off. It is mainly used to temporarily store running programs and data.
ROM, Read-Only Memory:It can only be read. The content is written during manufacture or through a special method. It is often used for firmware storage.
Flash Memory:Non-volatile, commonly found in SSDs, USB flash drives and memory cards.
High-speed cache memory (Cache):Located between the CPU and main memory to speed up data access.
Main technology
DRAM (Dynamic RAM)：Mainstream computer memory needs to constantly refresh the stored charge.
SRAM (Static RAM)：No refresh required, faster but more expensive, often used for caching.
DDR series:From DDR, DDR2, DDR3, DDR4 to DDR5, speed and bandwidth are continuously improved.
Application areas
PCs and laptops
Servers and Data Centers
Smartphones and Tablets
Embedded Systems and Industrial Control
Major brands
Samsung
SK Hynix
Micron
Nanya (Nanya Technology)
Kioxia (Kioxia, formerly Toshiba Memory)
NAND flash memory
Basic definition and working principle
NAND flash memory is a non-volatile storage technology, meaning it retains data even after power is removed. It stores bit data by capturing electrons in a transistor. Its core structure usually includes a floating gate or charge trap layer, which represents different logical values by changing its charge state.
Storage unit classification
According to the number of bits each storage unit (Cell) can hold, NAND can be divided into the following categories, with differences in density, cost and lifespan:
type
Bits per Cell
Write times (P/E Cycles)
advantage
shortcoming
SLC (Single-Level Cell)
1
About 50,000 - 100,000
Fastest, most durable, low error rate
Extremely high cost, low capacity
MLC (Multi-Level Cell)
2
About 3,000 - 10,000
Balancing performance and cost
Moderate durability
TLC (Triple-Level Cell)
3
Approximately 500 - 3,000
High cost-effectiveness, currently the core of the mainstream market
Slower writing speed and shorter lifespan
QLC (Quad-Level Cell)
4
Approximately 100 - 1,000
Extremely high capacity density and lowest price
The lowest efficiency and the narrowest lifespan
2D and 3D NAND Technology Evolution
2D NAND (Planar NAND):The memory cells are arranged horizontally on the surface of the silicon wafer. As the process shrinks to the limit, the interference (coupling effect) between units increases, making it difficult to further expand the capacity.
3D NAND (Vertical NAND):Stack storage units vertically. This is like switching from building a bungalow to building a skyscraper, which can significantly increase the storage density and provide hundreds of layers of stacking capabilities under the same area, effectively solving the scaling problem of 2D structures.
Common application scenarios
NAND flash memory is widely used in modern electronic products:
Solid State Drive (SSD):Used in personal computers, servers and data centers to replace traditional mechanical hard disks (HDD).
mobile device:UFS or eMMC storage built into smartphones and tablets.
Removable storage:USB flash drive, SD memory card and microSD card.
Embedded systems:Industrial automation equipment, vehicle information systems and IoT devices.
Key controller features
In order to extend the life of NAND and ensure data accuracy, the main control chip (Controller) performs the following key operations:
Wear Leveling:Ensure that all memory cells are evenly distributed with erase and write times to prevent premature damage to specific areas.
Error Correction Code (ECC):Correct bit errors caused by charge loss or read and write interference.
Bad Block Management:Mark and skip damaged units produced in production or in use.
Garbage Collection:Move valid data and clear out-of-date blocks to maintain write performance.
NOR flash memory
Technical definition and operating principle
NOR flash memory is a non-volatile storage technology that gets its name from the way its internal memory cells are arranged likeNOR logic gate. Unlike NAND, NOR flash memory has a parallel address bus, allowing the system to perform random access directly to any byte in the memory. This property enables it to work like a read-only memory (ROM).
Comparison with NAND flash memory
characteristic
NOR flash memory
NAND flash memory
Access method
Random Access
Sequence access (Page/Block Access)
Reading speed
extremely fast
medium
Write/erase speed
slow
quick
storage density
Low (usually under 1Gb)
Very high (up to several Tb)
unit cost
high
Low
interface
Parallel or SPI
Sequence (I/O)
Execute-In-Place (XIP) Features
This is the most critical advantage of NOR flash memory. With support for random access, the central processing unit (CPU) can execute code directly from NOR flash memory without first copying the code to random access memory (RAM). This not only reduces system boot time but also reduces RAM capacity requirements.
Technical classification
Parallel NOR：It adopts a traditional parallel interface with a large number of pins and high data transmission rate. It is often used in embedded systems that require extremely high performance.
Serial NOR (SPI NOR)：Using the Serial Peripheral Interface (SPI), with a very small number of pins and a small package size, it is currently the mainstream choice for wearable devices and IoT devices.
Application areas
Code storage:Used to store the microprocessor's firmware (Firmware) or operating system boot program (Bootloader), such as the computer's BIOS/UEFI.
Automotive electronics:Used in advanced driver assistance systems (ADAS) and digital instrument panels due to their fast start-up speed and high reliability.
Internet of Things (IoT) devices:Used to store core code for small devices.
Industrial control:For control units that require long-term stable operation and extremely low error tolerance.
key manufacturers
Currently, the world's major NOR flash memory suppliers include: Winbond, Macronix, GigaDevice and Micron.
magnetic random access memory
Physical mechanism and core architecture
Magnetic Random-Access Memory (MRAM) is a non-volatile memory that uses the spin properties of electrons to store data. Its core component is the Magnetic Tunnel Junction (MTJ), which consists of two ferromagnetic layers and a thin insulating layer (tunnel barrier) sandwiched between them.
Parallel status:Both magnetic layers are magnetized in the same direction and have low resistance, representing logic 0.
Anti-parallel state:The magnetization directions of the two magnetic layers are opposite and the resistance is high, representing logic 1.
Tunnel Magnetoresistive Effect (TMR):Data is read by detecting changes in resistance without moving charges, making reading extremely fast and using low power.
Technology evolution classification
As technology develops, MRAM has undergone major changes in the way it is written:
Toggle MRAM：The first-generation technology uses an external magnetic field to change the magnetization direction. The disadvantages are poor scalability and high current consumption.
STT-MRAM (spin transfer torque):A polarizing current is used to switch the magnetization direction of the free layer directly through the MTJ. It is currently the mainstream of commercialization, with high density and lower power consumption.
SOT-MRAM (Spin Orbit Moment):Separating the read and write paths, the write current does not pass through the tunnel barrier, which greatly improves the durability and switching speed, and is regarded as a potential replacement for SRAM.
Comparison with traditional memory
characteristic
SRAM
DRAM
NAND Flash
STT-MRAM
volatility
volatility
volatility
non-volatile
non-volatile
Reading and writing speed
extremely fast
medium
slow
quick
Durability (Cycles)
unlimited
unlimited
103 - 105
1012+
Static power consumption
High (leakage)
Medium (requires refresh)
extremely low
extremely low
unit area
Large (6T-8T)
Small (1T1C)
extremely small
Small (1T1J)
Application areas
Embedded Memory (eMRAM):Replace flash memory in microcontrollers (MCUs) and system-on-chips (SoCs) to improve performance and reduce standby power consumption.
Industry and Aerospace:It has high radiation resistance and high temperature resistance, and is suitable for critical data storage in extreme environments.
Edge AI:It is used for always-on devices because it has instant start-up characteristics and does not need to maintain refresh power.
Enterprise level cache:As the buffer memory of SSD, it provides power-off protection function to ensure that data will not be lost when power is interrupted.
Key advantages
Non-volatile:Data will not be lost after a power outage, making it truly "ready to use".
Unlimited durability potential:Compared to NAND, MRAM has almost no wear issues.
Low power architecture:Since no power is required to maintain the memory state, battery life can be significantly extended.
harddisk
Basic classification of hard drives
Mechanical hard disk (HDD): uses magnetic disks to store data, and relies on the movement of the magnetic head when reading and writing data.
Solid state drive (SSD): uses flash memory as the storage medium, requires no moving parts and is faster.
Hybrid hard drive (SSHD): combines the characteristics of HDD and SSD, storing commonly used data in the SSD part and other data in the HDD part.
Advantages and Disadvantages of Mechanical Hard Drives
Advantages:
Large capacity and low unit storage cost.
It has a relatively long service life and is suitable for large amounts of data storage.
Disadvantages:
Reading and writing speeds are slow.
Noise and vibration will be generated during operation.
High power consumption.
Applicable occasions
Suitable for backup and large-capacity data storage, such as file servers and video media storage.
It is not recommended to be used as a system disk because the boot speed and program execution performance are relatively low.
Shopping advice
Select the capacity according to your needs, commonly 1TB, 2TB and larger.
Pay attention to the hard drive speed (RPM). For example, 7200 RPM performs better than 5400 RPM.
Check the cache size (Cache). A larger cache can improve read and write efficiency.
Choose brands with good reputations, such as Seagate and Western Digital (WD).
SSD
What is SSD
SSD (Solid State Drive) is a storage device that uses flash memory as a storage medium. Compared with traditional mechanical hard disks (HDD), SSD has no moving parts, so it has faster access speeds, higher durability, and lower power consumption.
How SSD works
SSD mainly relies on NAND flash memory chips to store data, and manages the writing and reading of data through the controller. These controllers are responsible for functions such as data compression, error correction, and life-extending garbage collection.
Main features of SSD
High speed performance:SSDs provide faster boot times, data reading and writing speeds, and application loading times.
Durability:With no moving parts, SSDs are more resistant to shock and impact.
Quiet operation:Compared to traditional hard drives, SSDs operate almost silently.
Energy saving efficiency:SSD consumes less power and is suitable for laptops and mobile devices.
Types of SSD
SSDs can be divided into the following categories based on interface and shape:
SATA SSD：Uses SATA interface, suitable for most older computers.
M.2 SSD：It is thin and light, suitable for laptops and mainstream computers, and supports multiple protocols (such as SATA and NVMe).
NVMe SSD：Using PCIe channel, the speed is several times faster than SATA, suitable for applications requiring high performance.
SSD application scenarios
System acceleration:As a system disk, it greatly shortens boot and software startup time.
Game experience:Load game scenes quickly and reduce delays.
Professional use:Ideal for video editing, 3D rendering, and other high-performance tasks.
SSD purchasing advice
Confirm the interface type supported by your computer motherboard (such as SATA or NVMe).
Choose the appropriate capacity based on your needs, generally at least 512GB is recommended.
Pay attention to the product's read and write speed and durability (TBW or MTBF indicators).
Choose well-known brands to ensure quality and after-sales service.
FTL technology in solid state drives
What is FTL
FTL (Flash Translation Layer): It is a technology that implements logical address and physical address mapping in solid state drives (SSD).
Function: Convert the logical block address (LBA) issued by the host into the physical page address (PBA) of the NAND flash memory.
Goal: Improve SSD performance, life and reliability, and manage write operations.
Main functions of FTL
Address mapping: Corresponds the logical address on the host side to the actual stored physical address.
Wear Leveling: Ensure that each unit in the flash memory is used evenly to avoid premature aging.
Garbage Collection: Recycle invalid data pages to make space for writing.
Error Correction (ECC): Repairs errors during reading or writing to ensure data integrity.
FTL mapping method
Full mapping (Page-Level Mapping): Maps logical pages directly to physical pages, which is fast but requires more memory.
Block-Level Mapping: Mapping logical blocks to physical blocks, with low memory requirements but poor efficiency.
Hybrid Mapping: combines the advantages of full mapping and block mapping, taking into account efficiency and memory usage.
The impact of FTL on SSD performance
Effective FTL implementation can optimize read and write performance and reduce data latency.
Improve the life of NAND flash memory and delay wear and aging.
Affects the garbage collection efficiency of SSD, which is especially important for random write operations.
The future development direction of FTL
Improve the intelligence of the mapping algorithm and reduce the impact of garbage collection on performance.
Utilize larger-capacity DRAM or internal cache to optimize address query efficiency.
Design more targeted FTL structures for different application scenarios.
CD
definition
Optical Disc is a storage medium that uses laser reading and writing technology to access data. Optical discs read or burn data optically and can be used to store images, audio, software or other digital data.
Main types
CD（Compact Disc）: Capacity is about 700 MB, commonly used for music or data storage
DVD（Digital Versatile Disc / Digital Video Disc）: Capacity approximately 4.7 GB (single layer), 8.5 GB (double layer), used for video and data backup
Blu-ray Disc（BD）: Capacity of approximately 25 GB (single layer), 50 GB (double layer), supports high-definition video and large data storage
characteristic
Non-volatile storage: data is retained even without power
High compatibility: can be used on various optical disc reading devices
Durability: Resistant to magnetic field interference, but need to avoid scratches and exposure
Writability: Depending on the specification, it can be written once or rewritten multiple times.
application
Music and video distribution
Data backup and file storage
Software and system installation CD
Games & Entertainment Media
Major brands
Verbatim (Verbatim)
TDK
SONY
Maxell
Panasonic
Philips
DVD
definition
DVD, whose full name is "Digital Versatile Disc" or "Digital Video Disc", is an optical disc used to store digital data. It is mainly used for video playback, music storage and data backup.
Specifications and Capacity
Single sided single layer: 4.7 GB
Single sided double layer: 8.5 GB
Double-sided single layer: 9.4 GB
Double sided double layer: 17 GB
Main purpose
Video playback: such as movies, TV shows, etc.
Music storage: high-quality music albums.
Data Backup: Save files, photos and other data.
Features
It has a larger capacity than a CD and is suitable for storing high-quality movies and music.
Supports multiple formats, such as DVD-Video, DVD-Audio, etc.
Available in single or double sided, single or double layer designs.
Comparison with other media
Compared with CDs, DVDs have higher storage capacity and faster reading speeds; but compared with Blu-ray Discs (Blu-ray), DVDs have relatively lower capacity and image quality.
DVD-R
definition
DVD-R (Digital Versatile Disc Recordable) is a write-once optical disc that belongs to WORM (Write Once Read Many) media. Users can burn data, videos or audio to discs, but they cannot be deleted or overwritten once the writing is complete.
Main features
Capacity: Approximately 4.7 GB single layer, up to 8.5 GB dual layer
Writing speed: common 1x, 2x, 4x, 8x, 16x, etc.
High compatibility: can be read on most DVD players and recorders
Data cannot be overwritten: once written, it cannot be deleted or rewritten
Manufacturing process (abbreviated)
Base material injection molding (including guide groove pregroove)
Dye coating (to form a writable recording layer)
Metal reflective layer coating
Protective layer coating and lamination
Pre-write (test write)
Quality inspection and packaging for shipment
application
Data backup and file storage
Video recording and playback
Software or system installation CD
Major brands
Verbatim
TDK
SONY
Maxell
Panasonic
Philips
DVD Pre-write
definition
DVD Pre-write is inDVD-ROne of the final processes in the optical disc production process, it is usually carried out after the disc forming, coating, lamination and protective layer are completed. This step uses special equipment to perform limited pre-writing or test writing on the optical disc recording layer to check the optical disc recording quality and write necessary manufacturing identification information.
Process location
Base material injection molding (including guide groove pregroove)
Verify record layer consistency and write stability
Reduce the risk of consumer-side burning or reading failure
DVD rewritable disc
concept
DVD Rewritable, common specifications are DVD-RW and DVD+RW, is an optical disc that can be read and written multiple times. Unlike DVD-R, which is burned once, DVD-RW can erase original data and rewrite it, making it suitable for data backup and temporary storage.
Main features
capacity:Approximately 4.7 GB for single tier and up to 8.5 GB for dual tier.
Number of times that can be rewritten:Generally up to about 1,000 times.
compatibility:High compatibility with DVD players and DVD-ROMs, but some older devices may only support DVD-R.
speed:Commonly used are different speeds such as 1x, 2x, 4x, 6x, 8x, 16x, etc.
type
DVD-RW：Promoted by the DVD Forum, it uses phase change recording technology and can be written and erased repeatedly.
DVD+RW：Driven by the DVD+RW Alliance, write management is more refined and compatibility is gradually improved.
application
Data backup
Video recording and temporary storage
Software testing and file exchange
Temporary data storage
Major brands
Verbatim (Verbatim)
TDK
SONY
Maxell
Panasonic
Philips
BD
definition
BD, full name "Blu-ray Disc", is a high-density optical disc used to store high-definition videos, large-capacity data and high-quality music. It is named after using blue-violet laser for reading and writing.
Specifications and Capacity
Single layer: 25 GB
Dual layer: 50 GB
Tier 3: 100 GB (BD XL)
Four tiers: 128 GB (BD XL)
Main purpose
Video playback: Supports 4K UHD and 3D videos.
Data storage: suitable for high-capacity backup and file preservation.
Game media: used for next-generation game consoles, such as PlayStation.
Features
Supports high-definition images (1080p, 4K) and high-resolution sound effects.
Large capacity, suitable for storing long-term or high-quality content.
Compared with DVD, BD's storage capacity, image quality and sound effects have been greatly improved; however, due to higher technology costs, its price is usually higher than DVD.
