🚨 “Redis Cluster uses consistent hashing”
That’s what most people say.
But is it really true? 🤔
I was thinking about this recently…
In consistent hashing:
👉 Adding a node → minimal data movement
But in Redis:
👉 We have 16384 slots
👉 When a new node is added
👉 We rebalance slots across nodes
💡 Which means:
👉 A significant number of slots can move
👉 Sometimes from multiple nodes
So the question is 👇
❓ Is Redis really doing consistent hashing?
❓ Or is it just controlled sharding?
💡 Reality:
Redis Cluster uses:
👉 Hash Slots (not a ring)
👉 CRC16(key) % 16384 → slot → node
So what’s the difference?
🔹 Consistent Hashing → minimizes movement mathematically
🔹 Redis Slots → controls movement operationally
🧠 Real insight:
Redis is not optimizing for:
👉 “minimum movement”
It’s optimizing for:
👉 “manageable and predictable rebalancing”
✔ You can move slots gradually
✔ You can rebalance manually
✔ No full system rehash
👉 That’s why Redis chose slots over a ring
💡 Final thought:
👉 Consistent hashing = elegant theory
👉 Redis slots = practical engineering
Curious to hear your thoughts 👇
Do you consider Redis Cluster as consistent hashing or sharding?
Want to go deeper into System Design?
📸 Follow on Instagram for quick tech content:
Follow for more on AI, Java & System Design 🚀
#SystemDesign #Redis #ConsistentHashing #DistributedSystems #Scalability #BackendDevelopment #Java #Developers #Tech