Computer Science › Computer Architecture

SSD vs HDD

Storage types and their very different speeds.

Also known as: SSD, HDD, solid state drive, hard disk drive, NVMe

Persistent storage keeps data when the power is off. Two kinds dominate.

• HDD (hard disk drive): spinning magnetic platters and a moving read/write head. Cheap per gigabyte, but slow, especially for random access, because the head has to move.
• SSD (solid-state drive): flash memory with no moving parts. Much faster, with far lower latency, and more expensive per gigabyte. NVMe SSDs connect over a faster interface than older SATA ones.

How different they are

Rough orders of magnitude (real numbers vary a lot):

An HDD seek can be tens of thousands of times slower than memory (latency numbers). Sequential reads (large files read in order) are much faster than random ones on both.

Why it matters to developers

• Databases and indexes are designed around the cost of disk access: B-trees minimize how many reads a lookup needs (database indexes).
• Random vs sequential access. Appending to a log and scanning in order is fast. Jumping around is not, especially on HDDs.
• Disk is the usual bottleneck for heavy data work, and “disk I/O wait” shows up in monitoring (system monitoring tools).
• Cloud volumes come in different performance tiers, with IOPS and throughput limits you pay for.
• Durability: data not yet flushed from memory to disk can be lost in a crash. Databases use write-ahead logs for this (WAL).
• Wear: SSDs endure a limited number of writes, though modern ones last for years of normal use.
• Network and object storage (like S3) are another tier, with high capacity and higher latency (object storage).

Rule of thumb: keep hot data in memory or on SSD, and cold data on cheaper storage (storage lifecycle).