Objective 3.4 covers the devices that store data persistently — surviving a power-off, unlike the RAM covered in objective 3.3. This is one of the broadest objectives in Domain 3, spanning traditional spinning hard drives, modern solid-state drives and their various interfaces and form factors, RAID configurations for combining drives into more resilient or higher-performing arrays, and removable/optical media.
The throughline across this entire objective is a set of tradeoffs that repeat at every level: speed vs. cost, capacity vs. size, and performance vs. redundancy. Nearly every comparison question on this objective is really asking you to weigh one of these tradeoffs in a specific scenario.
A hard disk drive (HDD) stores data magnetically on spinning platters, read and written by a moving mechanical arm with a read/write head. HDDs remain relevant primarily because they offer the lowest cost per gigabyte of any common storage technology, making them well-suited for bulk storage of large amounts of data where raw speed isn't the top priority — backups, archival storage, media libraries, and budget-conscious bulk storage builds.
An HDD's spindle speed, measured in RPM (revolutions per minute), describes how fast the internal platters spin. Faster spindle speeds generally mean faster data access — the read/write head can reach the needed data sooner, and data passes under the head more quickly once located.
| Spindle Speed | Typical Use Case |
|---|---|
| 5,400 RPM | Budget laptops, external/portable drives — prioritizes lower cost, lower power draw, and quieter operation over speed |
| 7,200 RPM | Standard desktop drives — the most common balance of speed and cost for everyday use |
| 10,000 RPM / 15,000 RPM | Enterprise/server drives where higher performance justifies the added cost, heat, and noise |
Exam Angle
Higher spindle speed = faster performance, but also more heat, more noise, more power draw, and (often) a higher price point. If a scenario emphasizes battery life or quiet operation (a laptop), expect the answer to favor a lower spindle speed or a switch to SSD entirely. If a scenario emphasizes raw performance for a server or workstation, a higher spindle speed (or again, SSD) is the better fit.
HDDs are manufactured in two standard physical sizes, named for the diameter of the platters inside:
| Form Factor | Typical Use | Characteristics |
|---|---|---|
| 3.5-inch | Desktop computers, servers, external desktop storage enclosures | Larger; typically higher capacity and lower cost per GB; requires more power |
| 2.5-inch | Laptops, portable external drives, some space-constrained desktops | Smaller and lighter; lower power draw; typically lower max capacity than 3.5-inch |
Matching Form Factor to Device
This pairs directly with the DIMM/SODIMM logic from objective 3.3: 3.5-inch = desktop/server (more physical space available, less concern about power/weight), 2.5-inch = laptop (limited space, battery-powered, weight matters). A 2.5-inch drive can be installed in a desktop with the right mounting bracket/adapter, but a 3.5-inch drive will not physically fit in a standard laptop drive bay.
A solid-state drive (SSD) stores data on flash memory chips with no moving parts at all, in contrast to an HDD's spinning platters and mechanical arm. This fundamental difference gives SSDs dramatically faster read/write speeds, near-instant data access (no time wasted waiting for a physical arm to move into position), silent operation, lower power consumption, and much better resistance to physical shock — at a meaningfully higher cost per gigabyte than HDDs.
An SSD's interface determines how it physically and electrically connects to the rest of the system, which has a direct and major impact on its maximum possible speed.
Exam Angle — NVMe vs. PCIe
A subtle but testable distinction: PCIe is the physical/electrical bus, while NVMe is the protocol/language spoken over that bus, purpose-built for flash storage. You'll commonly see them referenced together ("NVMe PCIe SSD") precisely because NVMe drives use a PCIe connection — but PCIe itself is a more general-purpose bus used for many other types of expansion cards too, not just storage.
Common Pitfall — M.2 Slot ≠ Guaranteed NVMe Speed
Don't assume "M.2" automatically means "NVMe speed." M.2 is just a physical form factor/connector — it can carry either a SATA-based signal (capped at SATA speeds) or a PCIe/NVMe-based signal (much faster), and the motherboard's specific M.2 slot must support the protocol the drive uses. Always check both the drive's supported interface and the motherboard slot's supported interface before assuming compatibility or expected performance.
RAID (Redundant Array of Independent Disks) combines multiple physical drives into a single logical unit, used to improve performance, redundancy (fault tolerance), or both, depending on the specific RAID level chosen. The exam tests five RAID levels — each represents a different tradeoff between speed, usable capacity, and the ability to survive a drive failure without losing data.
Splits ("stripes") data evenly across all drives in the array, with each drive holding a different piece of every file. This allows read/write operations to happen in parallel across multiple drives simultaneously, maximizing performance and total usable capacity (100% of all drives' space is usable).
The critical tradeoff: there is zero redundancy. If any single drive in a RAID 0 array fails, all data across the entire array is lost, because each file is split across every drive with no duplicate copy anywhere.
Writes an identical, complete copy of all data to two (or more) drives simultaneously. If one drive fails, the other contains a fully intact, immediately usable copy of all data, with no data loss.
The tradeoff: usable capacity is cut in half, since every byte of data is duplicated rather than contributing to total storage.
Stripes data across all drives like RAID 0, but also calculates and distributes parity data across all the drives in the array (no single dedicated parity drive — parity is spread evenly across all members). If one drive fails, the missing data can be mathematically reconstructed from the remaining data and parity information.
The tradeoff: better capacity efficiency than RAID 1 (only one "drive's worth" of space is sacrificed to parity, regardless of array size), but write performance is slower than RAID 0 due to the overhead of calculating parity on every write.
Functions like RAID 5, but calculates and stores two independent sets of parity data instead of one. This allows the array to survive the simultaneous failure of any two drives without data loss — a meaningful improvement for larger arrays, where the odds of a second drive failing during the lengthy rebuild process after a first failure become a real concern.
A "nested" or hybrid RAID level: drives are first paired into RAID 1 mirrors, and then those mirrored pairs are striped together using RAID 0. The result combines RAID 1's redundancy with RAID 0's performance, making RAID 10 a favorite for environments needing both high speed and strong fault tolerance — most commonly database servers handling heavy, performance-sensitive transactional workloads.
The tradeoff: like RAID 1, you sacrifice half of total raw capacity to redundancy — RAID 10 buys its speed-plus-safety combination with the most drives and the most "wasted" capacity of any level covered here.
| RAID Level | Min. Drives | Fault Tolerance | Usable Capacity | Best For |
|---|---|---|---|---|
| RAID 0 | 2 | None | 100% | Max speed/capacity; no critical data |
| RAID 1 | 2 | 1 drive | 50% | Simple redundancy for critical small-scale data |
| RAID 5 | 3 | 1 drive | (n−1)/n | Balanced general-purpose server storage |
| RAID 6 | 4 | 2 drives | (n−2)/n | Larger arrays needing extra fault tolerance |
| RAID 10 | 4 | 1 per mirrored pair | 50% | High-performance + redundancy (databases) |
Exam Angle
The single most commonly tested fact in this section: RAID 0 has no fault tolerance whatsoever — despite being a "RAID" level, it offers zero redundancy and is purely about speed/capacity. Also expect direct minimum-drive-count questions ("What is the minimum number of drives for RAID 5?" → 3) and scenario questions matching a stated need (maximum speed, simple mirroring, balanced enterprise storage, double fault tolerance, database performance) to the correct RAID level.
A USB flash drive (thumb drive) is a small, portable solid-state storage device that connects via USB, used for quick file transport between systems, basic backups, and bootable installation media for operating systems. Flash drives have no moving parts, are extremely portable, and have become commodity-priced for common capacities.
Memory cards are small flash-storage cards primarily used in cameras, smartphones, tablets, drones, and other compact devices where a full-size drive or even a flash drive wouldn't fit or wouldn't be appropriate. Common formats include SD (Secure Digital) and its smaller variant microSD, widely used across consumer electronics. As a working technician, you may need to read these via a built-in card reader or an external USB card-reader adapter.
An optical drive reads (and, on writable models, writes) data using a laser to detect microscopic pits and lands on the reflective surface of a disc — CDs, DVDs, and Blu-ray discs. While largely superseded by flash-based storage and network/cloud-based software distribution in everyday consumer use, optical drives still appear in specific contexts: legacy software/media that only exists on disc, certain backup/archival workflows, and some business or institutional environments still standardized on physical media for distribution or compliance reasons.
| Media | Typical Capacity | Note |
|---|---|---|
| CD | ~700 MB | Oldest, lowest capacity of the three |
| DVD | ~4.7 GB (single layer) | Common for software/movie distribution historically |
| Blu-ray | ~25 GB (single layer) | Highest capacity; uses a shorter-wavelength laser than DVD/CD |
Final Exam Reminders
HDD spindle speed = higher RPM is faster but hotter, louder, and more power-hungry. 3.5" = desktop/server, 2.5" = laptop.
SATA SSD = bottlenecked by the SATA interface itself (~600MB/s), regardless of how fast the flash memory could otherwise be.
NVMe = a protocol built for flash, running over PCIe (the physical bus) — the two terms are related but not synonyms.
M.2 = a form factor that can carry either SATA or PCIe/NVMe signaling — don't assume M.2 automatically means NVMe speed.
RAID 0 = striping, max speed, ZERO fault tolerance — the exam's favorite trap answer for "which RAID level provides redundancy."
RAID 1 = mirroring, 50% usable capacity, survives 1 drive failure.
RAID 5 = striping + single distributed parity, 3-drive minimum, survives 1 failure.
RAID 6 = striping + double distributed parity, 4-drive minimum, survives 2 simultaneous failures.
RAID 10 = mirrored stripes, 4-drive minimum, best speed+redundancy combo, but only 50% usable capacity.