Post

What is RAID technology

What is RAID technology

Table of contents

Description

RAID (Redundant Array of Inexpensive Disks) uses two or more drives in parallel to improve speed and/or reliability. It offers different levels optimized for various needs. These levels aren’t standardized, so implementations may vary.

This is in contrast to the previous concept of highly reliable mainframe disk drives known as SLED (Single Large Expensive Disk).

Previous basic concepts

  • Data striping is a technique where data is split into chunks and spread across multiple disk drives. Instead of storing a whole file on one disk, the system divides it into blocks and writes each block to a different disk in sequence.
  • Disk mirroring is a data storage technique where identical copies of data are written to two or more disks simultaneously.
  • Parity bit is a simple error-checking tool used in data storage and transmission. It adds a small piece of extra information that helps verify whether the data is correct.

RAID 0: Striping

Requires a minimum of 2 drives to implement.

The data is subdivided (striped) and written on the data disks.

RAID 0 diagram RAID 0 diagram

AdvantagesDisadvantagesBest Uses
I/O performance is greatly improved by striped disk arrayNO fault-tolerantAny application requiring high bandwidth
No parity calculation overhead is involvedFailure of one drive will result in all data in an array being lostVideo Production and Editing
Very simple designNO for critical processImage Editing
Easy to implement  

RAID 1: Mirroring

Requires a minimum of 2 drives to implement.

RAID 1 diagram RAID 1 diagram

AdvantagesDisadvantagesBest Uses
Fast reads and reliable writeUses double the disk spaceAny application requiring very high availability
Full data backup—easy recovery after failureSoftware setups may not allow easy disk swap (Hardware implementation recommended)Accounting
Easy to set up and manage Payroll

RAID 2 : Bit-level striping + Dedicated Hamming code parity

Each data bit is stored across multiple data disks, while its Hamming Code error correction (ECC) is saved on separate ECC disks. During reading, the ECC checks for accuracy and can fix errors from a single faulty disk.

RAID 2 diagram RAID 2 diagram

AdvantagesDisadvantages
Fixes data errors on the flyNeeds many error-checking (ECC) disks for small data sizes
Very high data transfer speedsVery expensive, only worth it for extreme speed needs
Better efficiency with more data disksOverall transaction speed equals that of a single disk
Simpler controller design than RAID 3, 4, & 5Not used commercially. Not practical for real-world systems

RAID 3 : Byte-level striping + Dedicated parity

Requires at least 3 drives to implement.

The data block is subdivided (striped) and written on the data disks. Then Stripe parity is created during write operations, saved on a parity disk, and used during reads to verify or recover data.

RAID 3 diagram RAID 3 diagram

AdvantagesDisadvantagesBest Uses
Very fast read and write speedsSlower transaction rate, like a single diskAny application requiring high throughput
Keeps working well even if a disk failsComplex controller designVideo production & live streaming
Efficient use of disksHard to set up with software RAIDImage and video editing

RAID 4: Block-level striping + Dedicated parity

Require a minimum of 3 drives to implement.

Each data block is stored on a separate disk. Parity for blocks in the same position (rank) is created during writes, saved on a parity disk, and used during reads to verify or recover data.

RAID 4 diagram RAID 4 diagram

AdvantagesDisadvantages
Very fast read performanceComplex controller design
Efficient disk usage (low parity overhead)Slow write speed and poor write performance
High overall read throughputHard and slow to rebuild data after disk failure
 Reading a single block is as slow as reading from one disk

RAID 5: Block-level striping + Distributed parity

Require a minimum of 3 drives to implement.

RAID 5 uses block-level striping with distributed parity. Parity data is spread across all drives, allowing the system to keep working even if one drive fails. Read and write operations are shared across multiple disks for better performance.

RAID 5 diagram RAID 5 diagram

AdvantagesDisadvantagesBest Uses
Fastest read performanceMedium slowdown if a disk failsFile & application servers
Decent write speedVery complex controller designDatabase servers
Efficient disk usage (low parity overhead)Single block reads are as slow as one diskWeb, email, and news servers
Good overall data transfer rate Intranet servers, general-purpose RAID setups

RAID 6: Block-level striping + Double distributed parity

Requires a minimum of 4 drives to implement.

To protect against two simultaneous disk failures, RAID 6 uses two separate parity calculations, each based on a different algorithm. This dual-parity setup ensures stronger fault tolerance.

RAID 6 diagram RAID 6 diagram

AdvantagesDisadvantagesBest Uses
Dual parity offers extra fault toleranceComplex controller designFile & application servers
Data is striped across drives for efficient accessHigh overhead for parity calculationsDatabase servers
Protects against multiple bad block failuresSlower writes unless using specialized hardware (ASIC)Web and email servers
Reliable even in degraded modeNeeds at least two extra drives (N+2 setup)Intranet servers, mission-critical systems

[Extra] Some Nested RAIDs

RAID 10 (1+0): Speed + Redundancy

Requires a minimum of 4 drives to implement.

RAID 10 diagram RAID 10 diagram

RAID 50 (5+0): Parity Across stripes

Requires a minimum of 6 drives to implement.

RAID 50 diagram RAID 50 diagram

RAID 0+1: Striped mirrors

Requires a minimum of 4 drives to implement.

RAID 0 + 1 diagram RAID 0 + 1 diagram

Useful tools

This post is licensed under CC BY 4.0 by the author.