Sharding splits a large database into smaller pieces (shards) on different servers. Each shard holds a subset of data. Formula: shard = user_id % num_shards. Problem: "celebrity effect" causes hot spots.

Discussion

0 comments

0% completed

🔍Search

High-Level Design0/247

Fault Tolerance vs High Availability

Scalability — Deep Dive

Fault Tolerance Basics

Redundancy & Replication Basics

Graceful Degradation

Patterns for Data Consistency

Latency vs Throughput

Single Point of Failure (SPOF)

🤖Agentic AI0/28

Back to HLD

12/14

mediumPhase 6 · Scaling & Reliability 5 min read

Consistent Hashing

Advanced Content

Free users can access Foundations (Phase 1-4). Upgrade for Advanced content, quizzes, and unlimited access.

Upgrade to Pro

Interactive — Try it yourself!

How to use

1Type a user_id and click "Add User" — see which shard it goes to (hash % 4)

2Click "Add 100 Users" — watch even distribution across shards

3Hit "Celebrity Effect" — 80 users flood Shard 0 (hot spot!)

4Click "Reset" to start fresh

In a nutshell

Discussion

0 comments