Ensuring data availability and integrity is critical for any distributed system. Failures, whether due to hardware, software, or environmental factors, are inevitable. This lesson explores how data redundancy, mirroring, and backup/restoration strategies protect systems from media failures and ensure smooth recovery.

1. Data Redundancy

Data redundancy involves storing multiple copies of data across different locations to ensure availability and protect against data loss.

Techniques for Data Redundancy

1. Replication:

   • Copies of data are maintained on multiple nodes.
   • Used in distributed systems for high availability and fault tolerance.

2. RAID (Redundant Array of Independent Disks):

   • Combines multiple physical disks into a single logical unit.
   • Levels like RAID 1, RAID 5, and RAID 6 offer varying degrees of redundancy and performance.

3. Erasure Coding:

   • Data is split into fragments and encoded with redundancy.
   • Only a subset of fragments is needed to reconstruct the data, reducing storage overhead compared to full replication.

2. Mirroring

Mirroring is a specific form of redundancy where an exact copy of data is maintained in real-time.

How Mirroring Works

• Every write to the primary disk or node is instantly duplicated on the mirror.
• If the primary fails, the mirror takes over seamlessly.

Advantages:

• Immediate failover ensures minimal downtime.
• High data integrity since the mirror is always up-to-date.

Disadvantages:

• Doubles storage costs due to the need for a full copy.
• May increase write latency for real-time updates.

Example:

• A banking system mirrors its transaction database between two data centers to ensure no data is lost during failures.

3. Media Failure Recovery

Media failures occur when storage devices (e.g., hard drives or SSDs) malfunction. Recovery from such failures requires proactive and reactive strategies.

Types of Media Failures:

1. Physical Damage: Disk corruption, broken hardware.
2. Logical Corruption: Software bugs or malware causing data errors.

Strategies for Media Failure Recovery

1. Hot Swapping: Replace failed drives in RAID arrays without shutting down the system.
2. Regular Health Monitoring: Tools like SMART (Self-Monitoring, Analysis, and Reporting Technology) to detect potential drive failures early.
3. Data Scrubbing: Periodically check and repair corrupted data blocks.
4. Data Replication: Store multiple copies of data across different nodes or regions to ensure quick recovery from failures.

4. Backup and Restoration

Backup is the process of creating additional copies of data for protection against catastrophic failures. Restoration involves recovering from backups after a failure.

Types of Backups

1. Full Backup:

   • A complete copy of all data.
   • Pros: Simplifies restoration.
   • Cons: Time-consuming and storage-intensive.

2. Incremental Backup:

   • Captures only the changes made since the last backup.
   • Pros: Saves storage and time.
   • Cons: Restoration may require multiple backups.

3. Differential Backup:

   • Captures changes since the last full backup.
   • Pros: Faster restoration compared to incremental backups.
   • Cons: Requires more storage than incremental backups.

Best Practices for Backup and Restoration

1. Backup Frequency: Schedule backups based on system criticality (e.g., daily for business-critical systems).

2. Offsite Storage: Store backups in a remote location or cloud to protect against regional disasters.

3. Testing Restorations: Regularly verify that backups can be restored to ensure they are usable.

4. Versioning: Keep multiple backup versions to recover from data corruption or ransomware attacks.

Example Scenario

A hospital’s patient records are backed up daily using:

• Full backup on weekends for complete redundancy.
• Incremental backups on weekdays to save storage.
• The backups are stored both locally and in the cloud to handle regional failures.