Questions
8 of 11
You need to cache data. Some data expires in 1 minute, some in 1 hour. How do you design this using OOP?
Difficulty: 3/10
Policy-Based Cache Design
I would hide the caching implementation behind an ICache abstraction and represent expiration as a cache policy rather than scattering TTL values throughout the application. Different consumers can provide different policies. The cache implementation can then enforce absolute or sliding expiration consistently.
javascript
- 1
Separate cache abstraction from cache implementation.
- 2
Represent TTL and other behavior as explicit policies.
- 3
Support absolute and, where appropriate, sliding expiration.
- 4
Define cache invalidation rules and consistency expectations.
- 5
For distributed applications, consider a distributed cache rather than process-local memory.
- 6
Monitor hit rate, eviction rate, memory usage, and latency.
Follow-up Questions
- What is cache-aside?
- How would you invalidate cached data after an update?
- When would you use distributed caching?