Storage: SSDs and Hard Drives Explained Clearly

RAM forgets its contents when normal power disappears. A computer therefore needs non-volatile storage to preserve the operating system, applications, documents, and media.

Two major storage technologies are solid-state drives and hard disk drives.

An SSD stores data electronically in flash memory. A hard drive stores data magnetically on rotating platters and uses a moving head to access it.

Both retain files, but their physical designs create very different performance and durability characteristics.

How a hard drive stores data

A hard disk drive, or HDD, contains one or more rigid platters coated with magnetic material. A motor spins them at high speed.

An actuator positions read-write heads extremely close to the platter surfaces. The heads detect or change magnetic patterns representing data.

To read a requested block:

1. The actuator moves the head to the correct track.
2. The drive waits for the desired sector to rotate under the head.
3. Electronics decode the magnetic signal.
4. Data moves through the drive controller to the computer.

The physical movement creates seek time and rotational delay.

Once positioned, a hard drive can transfer a large sequential region reasonably well. It struggles more with many small requests scattered across the disk because the head must keep moving.

How an SSD stores data

A solid-state drive uses flash-memory cells that store electrical charge states. It has no spinning platter or mechanical head.

An SSD controller manages:

• Mapping logical blocks to physical flash
• Error correction
• Wear leveling
• Caching
• Garbage collection
• Failed-cell handling

Flash memory is organized into pages and erase blocks. Data can be read and programmed at one granularity, but erasure happens in larger units. Updating data may require writing a new location and later reclaiming old blocks.

The controller hides much of this complexity and presents a normal block-storage interface to the operating system.

Why SSDs feel faster

An SSD can access scattered data without mechanically repositioning a head. This gives it much lower random-access latency.

Operating-system startup and application launch involve many files and small reads. An SSD can retrieve those pieces quickly, making the complete computer feel more responsive.

The performance difference is less dramatic for one long sequential transfer, especially when comparing a fast hard drive with an entry-level SSD over a limited interface. The SSD still offers other advantages:

• Silent operation
• Lower shock sensitivity
• Smaller form factors
• Often lower power use
• High random input-output performance

SATA and NVMe are interfaces, not storage materials

The words SATA and NVMe are often confused with physical drive type.

SATA is an older storage interface designed during the hard-drive era. SATA SSDs use flash memory but communicate through the SATA protocol and its speed limits.

NVMe is a protocol designed for non-volatile memory over PCI Express. It supports many parallel command queues and much higher potential throughput.

Common forms include:

• 2.5-inch SATA SSD
• M.2 SATA SSD
• M.2 NVMe SSD
• PCIe expansion-card SSD

M.2 describes a physical form factor and connector family, not automatically NVMe. An M.2 device can use SATA or PCIe depending on the drive and slot.

Compatibility requires checking both shape and protocol.

A concrete workload comparison

Imagine two tasks.

Copying one 20 GB video

Both drives perform mostly sequential reading and writing. The SSD is likely faster, but the difference depends on drive models, interfaces, caches, and destination speed.

Starting an operating system

The system reads many scattered files, metadata records, libraries, and configuration values. The hard drive repeatedly seeks to new locations. The SSD accesses them with far lower mechanical delay.

The second task often shows a much larger visible advantage.

This is why a single sequential speed benchmark cannot predict the entire experience.

Flash endurance and wear leveling

Flash cells tolerate a finite number of program-erase cycles. SSDs manage this limitation through wear leveling, spreading writes across available cells instead of repeatedly using the same physical area.

Manufacturers commonly specify endurance using measures such as terabytes written.

Ordinary consumer workloads often remain well within rated endurance, but heavy write-intensive applications should choose appropriate drives and monitor health.

SSDs also reserve spare capacity that the controller can use for replacement blocks and internal management. Keeping some free space can help garbage collection and sustained write performance.

Caches and sustained speed

Drives use caches to improve performance.

Some SSDs temporarily treat a portion of flash as a faster single-bit-per-cell cache. Short writes can be very fast. During a long transfer, the cache may fill, and sustained speed drops to the underlying flash rate.

Hard drives also use memory caches, but they cannot eliminate mechanical seek costs.

When evaluating a drive for large continuous writes, inspect sustained performance after caches fill rather than relying only on the peak number.

Failure modes

Any storage device can fail.

Hard drives may experience:

• Mechanical wear
• Motor or actuator failure
• Head crashes
• Damaged sectors
• Shock damage

SSDs may experience:

• Flash wear
• Controller failure
• Firmware defects
• Electrical damage
• Uncorrectable cells

SSDs can sometimes fail with less audible warning because no mechanical noise announces a problem. Health-monitoring data can help, but it cannot predict every failure.

The correct response is not to find an immortal drive. It is to maintain backups.

Deletion and data recovery

Deleting a file usually removes or updates file-system references; it does not necessarily erase every underlying physical representation immediately.

On SSDs, the operating system can send a TRIM command indicating which blocks are no longer needed. The controller may erase or reuse them later, making recovery less predictable.

Secure disposal requires methods appropriate to the device:

• Built-in secure erase or sanitize commands
• Cryptographic erase when full-disk encryption was used correctly
• Physical destruction for high-risk requirements

Repeated overwriting practices designed for magnetic disks do not map neatly onto flash translation and wear leveling.

Storage is not a backup

A second folder on the same drive is not a backup. If the drive fails, is stolen, or is encrypted by malware, both copies can disappear.

A robust backup strategy keeps independent copies, ideally including:

• Multiple copies
• More than one storage medium or system
• At least one copy separated from the primary device
• Tested restoration procedures

Cloud synchronization can be useful, but synchronized deletion or corruption may propagate. Check whether the service provides version history and recovery.

Choosing between SSD and HDD

SSDs are usually best for:

• Operating systems
• Applications
• Portable computers
• Databases and active projects
• Workloads with random access

Hard drives remain useful for:

• Large economical capacity
• Media collections
• Local backup targets
• Sequential archival workloads

Many systems combine them: SSD for active work and HDD or network storage for large, less-frequently accessed data.

Common misunderstandings

"An SSD has no moving parts, so it cannot fail"

It avoids mechanical failure modes but can still fail through controller, flash, firmware, electrical, or manufacturing problems.

"M.2 always means NVMe"

M.2 is a form factor. The device may communicate using SATA or PCIe/NVMe.

"A larger advertised sequential speed makes every task proportionally faster"

Small random operations, latency, software behavior, and other bottlenecks can dominate.

"Keeping files on a reliable drive removes the need for backup"

All drives can fail, and hardware reliability does not protect against deletion, theft, malware, or disaster.

Knowledge check

1. Why are SSDs much faster at scattered reads?

They access flash electronically without moving a mechanical head to each requested location.

2. What is the difference between M.2 and NVMe?

M.2 describes a physical form factor and connector. NVMe is a storage protocol commonly carried over PCI Express.

3. Why can an SSD slow during a very long write?

Its fast cache may fill, exposing the slower sustained write rate of the underlying flash and controller cleanup work.

4. Why is a second copy on the same drive not a sufficient backup?

Both copies share the same failure, theft, malware, and physical-damage risks.

The one idea to remember

SSDs and hard drives both preserve data, but their physical access mechanisms shape latency, throughput, endurance, noise, power use, and failure behavior.

Choose storage for the workload, and protect important data with independent, tested backups regardless of the drive type.

Next, we will study the motherboard and the communication links that turn separate components into one functioning system.