Module: Advanced Operating Systems

Instructor: Dr Julius Raphael

Topic: Redundant Array of Independent
Disks (RAID) Concepts

RAID is a technology used in computer storage systems to combine multiple physical disks
into a single logical unit. This approach is designed to improve performance, redundancy,
or both. RAID is widely used in servers, data centers, and performance-critical systems to
ensure data reliability and availability.

Key Objectives of RAID

1. Performance: Enhance read/write speeds by splitting or replicating data across
multiple disks.
2. Redundancy: Protect data against disk failures by introducing fault-tolerant
mechanisms.
3. Efficiency: Optimize storage utilization depending on the RAID level.

Core Concepts of RAID

To understand RAID, it is essential to grasp four primary concepts:

1. Striping

 Definition: Striping involves dividing data into smaller chunks and writing them
across multiple disks sequentially. Each chunk is stored on a separate disk, which
enhances read and write performance by allowing multiple disks to operate
simultaneously.
 Advantages: Improves performance as multiple disks handle data requests in parallel.
 Disadvantages: Provides no redundancy; if a single disk fails, all data is lost.

2. Mirroring

 Definition: Mirroring creates identical copies of data on two or more disks. Each disk
holds a complete replica of the data, ensuring fault tolerance.
 Advantages: High data redundancy; if one disk fails, the mirrored disk still contains
the data.
 Disadvantages: Doubles the storage requirement, as each piece of data is duplicated.

1
3. Parity

 Definition: Parity is a technique used to provide redundancy by calculating

checksums of data blocks. These checksums are stored across disks and can be used
to reconstruct lost data in the event of a disk failure.
 Advantages: Offers fault tolerance without requiring full duplication like mirroring.
 Disadvantages: Rebuilding data from parity can be time-consuming and impacts
performance.

4. Checksum

 Definition: A checksum is a value derived from a data block using a specific

algorithm, such as cyclic redundancy check (CRC). It is used to detect errors and
verify data integrity during transmission or storage.
 Advantages: Ensures data consistency and detects corruption or unintended
modifications.
 Disadvantages: Requires additional computation, which may affect performance in
some systems.

RAID Levels and Their Characteristics

1. RAID 0 (Striping – No Redundancy)

 Description: Data is split evenly across two or more disks. This level focuses on
performance but lacks redundancy. If one disk fails, all data is lost.
 Use Case: High-speed applications where data loss is acceptable (e.g., temporary
processing).

Illustration:

Disk 1: A1 A3 A5 A7
Disk 2: A2 A4 A6 A8

2. RAID 1 (Mirroring)

 Description: Data is duplicated (mirrored) across two disks. Provides high reliability
as a failure of one disk does not result in data loss.
 Use Case: Critical systems requiring high redundancy (e.g., financial databases).

Illustration:

Disk 1: A1 A2 A3 A4
Disk 2: A1 A2 A3 A4 (Mirror of Disk 1)

3. RAID 5 (Striping with Parity)

2
  Description: Combines striping with parity. Data and parity (used for recovery) are
distributed across disks. Can tolerate one disk failure.
 Use Case: Balanced performance and fault tolerance for storage servers.

Illustration:

Disk 1: A1 A2 Parity
Disk 2: A3 Parity A4
Disk 3: Parity A5 A6

4. RAID 6 (Striping with Double Parity)

 Description: Similar to RAID 5 but uses two parity blocks for enhanced redundancy.
Can tolerate two simultaneous disk failures.
 Use Case: Large-scale systems requiring high fault tolerance (e.g., enterprise storage
arrays).

Illustration:

Disk 1: A1 A2 Parity1 Parity2

Disk 2: A3 Parity1 Parity2 A4
Disk 3: Parity1 A5 A6 Parity2
Disk 4: Parity2 A7 A8 Parity1

5. RAID 10 (RAID 1 + RAID 0)

 Description: RAID 10, also known as RAID 1+0, combines the advantages of RAID
1 (mirroring) and RAID 0 (striping). Data is first striped across multiple disks to
improve performance and then mirrored to ensure redundancy. This configuration
provides both high performance and fault tolerance, making it ideal for critical
applications.
 How It Works:
1. Data is divided into blocks and striped across two or more disks.
2. Each striped block is then mirrored onto another set of disks, creating
complete redundancy.
3. If a disk in one mirror pair fails, the system continues to operate using the
mirrored disk.
 Key Features:
o Performance: Striping ensures faster read and write speeds since multiple
disks are used simultaneously.
o Redundancy: Mirroring ensures that data is duplicated, offering protection
against disk failures.
o Scalability: Can scale with the number of disks, but requires at least four
disks to implement.
3
  Advantages:
o High performance due to striping.
o High fault tolerance due to mirroring.
o Simpler recovery process compared to parity-based RAID levels.
 Disadvantages:
o High cost: Requires twice as many disks as the usable storage capacity.
o Limited storage efficiency (only 50% of total disk capacity is usable).
 Use Case: RAID 10 is ideal for high-performance databases, applications requiring
minimal downtime, and environments where both performance and reliability are
critical.

Illustration:

Mirror 1: Disk 1: A1 A3 A5 A7
Disk 2: A1 A3 A5 A7

Mirror 2: Disk 3: B1 B3 B5 B7
Disk 4: B1 B3 B5 B7

 Description: Combines the striping of RAID 0 with the mirroring of RAID 1. This
level provides both high performance and fault tolerance.
 Use Case: High-performance databases and applications requiring redundancy.

Illustration:

Mirror 1: Disk 1: A1 A3 A5 A7
Disk 2: A1 A3 A5 A7

Mirror 2: Disk 3: B1 B3 B5 B7
Disk 4: B1 B3 B5 B7

Advantages of RAID
 Improved Performance: Enhanced read/write speeds with RAID 0, 5, 6, and 10.
 Fault Tolerance: Redundancy ensures data availability even during disk failures
(RAID 1, 5, 6, and 10).
 Scalability: Supports large storage systems by aggregating multiple disks.

Disadvantages of RAID
 Cost: Higher RAID levels (e.g., RAID 10) require more disks, increasing costs.
 Complexity: Configurations such as RAID 6 and RAID 10 may require specialized
hardware or expertise.
 Rebuild Time: Rebuilding data from parity in RAID 5 and RAID 6 can be slow,
impacting performance.

4
Hardware vs. Software RAID
1. Hardware RAID:
o Managed by dedicated RAID controllers.
o Provides better performance and independence from the operating system.
2. Software RAID:
o Managed by the operating system.
o Cost-effective but uses system resources, which can impact performance.

RAID in Practice
RAID is crucial for:

 Servers and Data Centers: Ensures availability and performance.

 Enterprise Applications: Supports critical systems like databases and ERP.
 Personal Use: Provides performance boosts for gaming or media editing.