Understanding Server Storage Types: HDD, SSD, RAID & Network Storage Explained
This article explains the different server storage types—including mechanical HDDs, solid‑state SSDs, various interfaces, RAID arrays, and network‑attached storage—while covering their structures, performance characteristics, Linux device handling, and how to choose the right solution based on cost, capacity, speed, and reliability.
Classification by Storage Medium
Disks provide persistent storage and are divided into two main categories: mechanical hard disk drives (HDD) and solid‑state drives (SSD).
Mechanical Disk (HDD)
HDDs consist of multiple platters with two‑sided storage. Data is organized into tracks, cylinders, and sectors (typically 512 bytes). Accessing data requires moving the read/write head to the correct track (seek) and waiting for the platter to rotate, making random I/O slower than sequential I/O.
Common techniques to improve HDD performance include:
Caching to eliminate I/O latency
Filesystem layout and write‑ahead logging (WAL)
Distributing workloads across different disks
Using outer tracks for higher throughput
Optimized seek algorithms
Solid‑State Disk (SSD)
SSDs use flash memory with no moving parts, offering faster read/write speeds and lower latency for both sequential and random I/O. However, SSDs still suffer from erase‑before‑write constraints, causing garbage collection overhead that makes random I/O slower than sequential I/O.
Minimum read/write units differ:
HDD: sector, usually 512 bytes
SSD: page, typically 4 KB or 8 KB
Classification by Interface
ATA/IDE
SCSI
SAS (Serial Attached SCSI)
SATA (Serial ATA)
FC (Fibre Channel)
NVMe (Non‑Volatile Memory Express)
Storage Arrays
Multiple disks can be combined into a storage array (RAID) to increase capacity, performance, and reliability. Common RAID levels (RAID 0, 1, 5, 6, etc.) offer different trade‑offs between cost, performance, and availability.
Network‑Attached Storage
Disks can be grouped into a network storage cluster and exposed via protocols such as NFS, SMB/CIFS, or iSCSI. Performance may be affected by network congestion, so client‑side analysis of load and I/O latency is important.
Disks as Block Devices in Linux
In Linux, a disk appears as a block device (e.g.,
/dev/sda) with a major and minor number. Partitions are represented as
/dev/sda1,
/dev/sda2, etc.
Summary
Server storage types fall into three categories: individual disks (HDD or SSD), storage arrays (RAID), and network‑attached storage. Choose based on price, capacity, performance, and stability—SSD for speed, HDD for large cheap capacity, RAID for reliability and throughput, and network storage for shared access.
Next Topic Preview
Disk Performance Inspection
When facing I/O performance issues, consider the following questions:
What is the system’s IOPS and each disk’s IOPS?
What is the throughput of the system and each disk?
What is the disk utilization?
Which application or user is accessing the disk?
Which filesystem or file is being accessed?
Ops Development Stories
Maintained by a like‑minded team, covering both operations and development. Topics span Linux ops, DevOps toolchain, Kubernetes containerization, monitoring, log collection, network security, and Python or Go development. Team members: Qiao Ke, wanger, Dong Ge, Su Xin, Hua Zai, Zheng Ge, Teacher Xia.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.