Cache_As_Sor.md

Cache-As-SOR

• Motivation
• Solution
  • Concepts
  • Sub-Patterns
  • Implementation
• Pros & Cons
• Consideration
• When To Use
• References

Motivation

Solution

Concepts

Sub-Patterns

Pattern Name	Diagram	Description
Read-Through		If requested data is not in cache, cache will load the data from database and let application read synchronously.
Write-Through		After application writes data to cache, cache synchronously writes the same data to database.
Write-Behind		After application writes data to cache, cache asynchronously writes the same data to database (When satisfy a certain criteria).
Refresh-Ahead		Cache automatically and asynchronously reloads (refreshs) any recently accessed cache entry from database before its expiration.

Implementation

Pros & Cons

• Write-Through
  • Pros
    • No risk of data loss.
  • Cons
    • Slow (cache can only ack back until writing data to database successfully).
    • When a new cache node is added, the new cache node will be empty until there is a update from the application.
• Write-Behind
  • Pros
    • Faster (cache can ack back without writing data to database).
  • Cons
    • Risk of data loss: Data may lose after cache goes down, but before it have been written to database.
    • More complex to implement.
• Refresh-ahead
  • Pros
    • Faster
  • Cons
    • Not accurately predicting which items are likely to be needed in the future can result in reduced performance than without refresh-ahead.

Consideration

Topic	Consideration	Possible Solution Options

When To Use

References