💬
✉️
📸
▶️
Loading…
Loading…
Open SystemDecoder on a larger screen to build systems, run simulations, and inject chaos.
Daily Challenge
SystemArena
5
Questions
Think Deep
1×
Daily Shot
5 sharp system design questions. One attempt daily — streaks, XP & leaderboard.
What's waiting for you on desktop
Live Simulations
Watch latency spike, queues fill, and nodes fail in real time. Every slider change is instant.
Visual Architecture Canvas
Drag nodes, draw edges, and build any distributed system topology from scratch.
Chaos Engineering
Kill servers, introduce packet loss, throttle CPUs — and watch your system react.
Real-time Insights
Throughput, p99 latency, error rates — all charted live as your simulation runs.
40+
Concepts
<1s
Feedback
∞
Replays
"The best way to understand a distributed system
is to break it."
Google Search
🔍
8.5 billion queries/day. Results in <200ms.
Crawl pipeline → Index pipeline → Query-serving pipeline
Request Flow
The scale
Google processes ~8.5 billion queries per day. Every query must return ranked, personalised results in under 200ms including network round-trips.
Three decoupled pipelines
Google Search is three pipelines running at completely different timescales: (1) the + + query-serving pipeline (<200ms); (2) the web crawl + build pipeline (hours/days); and (3) the + real-time suggestion pipeline (<20ms).
Why decoupled?
The crawl pipeline can be completely down for hours without affecting the query-serving pipeline. This decoupling is achieved via Kafka and hot-swappable index segments.
Key Insight
Google Search is not one system — it's three pipelines running at vastly different timescales, decoupled by design so each can fail, scale, and deploy independently.
Overview