What is LFU in Redis: Explanation and Usage

In Redis, LFU stands for "Least Frequently Used" and is a cache eviction policy that removes keys used least often when memory is full. It helps keep frequently accessed data in memory by tracking usage frequency and evicting the least used keys first.

How It Works

Imagine you have a small shelf where you keep your favorite books, but the shelf can only hold a limited number. When the shelf is full and you want to add a new book, you decide to remove the book you read the least often. This is similar to how Redis's LFU eviction policy works.

Redis tracks how often each key is accessed. When memory is full, it removes the keys that have been used the least frequently to make space for new data. This way, the most popular or frequently used data stays in memory, improving performance.

LFU is different from other policies like LRU (Least Recently Used), which removes keys based on how recently they were accessed rather than how often. LFU focuses on long-term usage frequency, making it useful when you want to keep data that is consistently popular.

Example

This example shows how to configure Redis to use the LFU eviction policy and how it affects key eviction.

redis

CONFIG SET maxmemory 100kb
CONFIG SET maxmemory-policy allkeys-lfu

SET key1 "value1"
SET key2 "value2"

# Access key1 multiple times to increase its frequency
GET key1
GET key1
GET key1

# Add many keys to fill memory and trigger eviction
for i in {3..1000}; do redis-cli SET key$i "value$i"; done

# Check if key1 still exists (likely yes because it is frequently used)
EXISTS key1
# Check if key2 still exists (likely no because it was used less)
EXISTS key2

Output

OK
OK
OK
"value1"
"value1"
"value1"
OK
(integer) 1
(integer) 0

When to Use

Use LFU eviction in Redis when you want to keep data that is frequently accessed over time, even if it was not accessed recently. This is helpful for caching scenarios where some data is consistently popular, like user session data, product information, or frequently requested API results.

LFU is ideal when you want to avoid evicting data that is important but might not have been accessed very recently. It helps maintain a cache that favors long-term popularity rather than just recent usage.

Key Points

• LFU stands for Least Frequently Used eviction policy in Redis.
• It removes keys that are used least often when memory is full.
• LFU tracks usage frequency, not just recent access.
• Useful for caching data with consistent popularity over time.
• Configured with maxmemory-policy allkeys-lfu in Redis.

Key Takeaways

LFU in Redis evicts the least frequently used keys to keep popular data in memory.

It tracks how often keys are accessed, not just how recently.

Use LFU for caches where long-term popularity matters more than recent use.

Configure LFU with the maxmemory-policy setting in Redis.

LFU helps optimize memory by preserving frequently accessed data.