Will the Guru

Definition

• Once the data is successfully submitted to the system, it is not lost.

How to support durability?

1. Backup
   • Full backup
     • Pros: Short restoration time
     • Cons: Long creation time
   • Differential backup
     • Pros: Smaller than full backup, shorter creation time
     • Cons: Longer restoration time than full backup
   • Incremental backup
     • Full + Differential
     • Pros: Small size and short creation time
     • Cons: Long restoration time and more complex restoration process

2. RAID

• RAID is focused on local disk redundancy and is ideal for protecting against individual disk failures. It offers different levels of protection and performance optimization but does not protect against site-wide disasters.

3. Replication

• Replication is focused on maintaining copies of data in multiple locations, which can be local or geographically dispersed. It is suitable for disaster recovery and ensuring data availability across different sites but involves more overhead and complexity than RAID.

Checksum

• On writing data, checksum will be generated and this checksum will be compared on reading data
  • This ensures the data is not corrupted
  • (Retrieved data checksum vs Original checksum)

Note

• Availability != Durability
  • Availability
    • About system uptime
    • “Can I access my data right now?”
  • Durability
    • About storing data without losing it
    • “Will my data still be there in the future?”

Reference

• https://leetcode.com/explore/interview/card/system-design-for-interviews-and-beyond/722/how-to-define-system-requirements/4745/