Cache-Aside

Cache-Aside

Let the application own cache hits, source fallback, cache population, and post-write invalidation.

About 20 minFoundationReviewed Aug 2026Progress unavailableRead to the end

What you will learn

Let the application own cache hits, source fallback, cache population, and post-write invalidation.

The application owns the miss path#

Cache-aside is a common read pattern: check the cache, read the source on a miss, then populate the cache. The database remains the source of truth and only requested values enter the cache.

Cache-aside hit and miss paths

The application checks key user colon 42. A hit returns directly. A miss causes a database read followed by cache population before the response.

value = cache.get("user:42")
if value is missing:
    value = database.read_user(42)
    cache.set("user:42", value, ttl)
return value

Writes need a separate decision#

Worked example

Change Maya to May

1. UPDATE users SET name = 'May' WHERE id = 42
2. DELETE cache key user:42
3. Next read reloads the committed value

Updating the source and then deleting the cached copy is often simpler than updating every cached representation. It still has a failure window: the database can commit while deletion fails. No ordering is universally safe; state the consistency requirement and recovery mechanism.

Interviewer follow-ups

Use these prompts to extend the design. The useful answer is a reasoned change, not a memorized sentence.

What happens if two requests miss simultaneously?
What if the database update succeeds but cache deletion fails?
Should a not-found result be cached?
Can callers bypass the cache for a read-after-write?

Apply the lesson

What you will learn

Let the application own cache hits, source fallback, cache population, and post-write invalidation.

The application owns the miss path#

Cache-aside is a common read pattern: check the cache, read the source on a miss, then populate the cache. The database remains the source of truth and only requested values enter the cache.

Cache-aside hit and miss paths

The application checks key user colon 42. A hit returns directly. A miss causes a database read followed by cache population before the response.

value = cache.get("user:42") if value is missing: value = database.read_user(42) cache.set("user:42", value, ttl) return value

Writes need a separate decision#

Worked example

Change Maya to May

1. UPDATE users SET name = 'May' WHERE id = 42
2. DELETE cache key user:42
3. Next read reloads the committed value

Interviewer follow-ups

Use these prompts to extend the design. The useful answer is a reasoned change, not a memorized sentence.

What happens if two requests miss simultaneously?
What if the database update succeeds but cache deletion fails?
Should a not-found result be cached?
Can callers bypass the cache for a read-after-write?

Apply the lesson

What you will learn

The application owns the miss path#

Writes need a separate decision#

Change Maya to May

Related practice problems

Finish this lesson

Cache-Aside

What you will learn

The application owns the miss path#

Writes need a separate decision#

Change Maya to May

Related practice problems

Finish this lesson