System Design Atlas

14 Google-style problems, interview-ready answers.

Each solution covers requirements, capacity math, architecture, deep dives, and failure modes — sized for a 60-minute answer with earned-secret bonuses you can reach for.

14Problems

6Categories

11Patterns

Browse the atlas

Networking & Traffic Management

2 problems

Design an L4 (Transport-Layer) Load Balancer

Route TCP and UDP flows to healthy backends while preserving per-flow stickiness and keeping the data path extremely fast.

Stateful hot pathAvailabilityPartitioningObservabilityCost / efficiency

Design a Distributed Rate Limiter

Enforce per-user, per-tenant, and per-IP quotas at gateway scale without turning the limiter itself into the bottleneck.

ConsistencyStateful hot pathPartitioningAvailabilitySafety / correctness

Storage & Query Systems

2 problems

Lexicographic Range Query Store

Build an ordered store that supports inclusive lexicographic range scans across billions of keys.

PartitioningConsistencyStreamingCost / efficiency

Design a Distributed Storage System (Object Store)

Choose block, file, or object storage, then justify durability, placement, repair, and metadata consistency.

ConsistencyPartitioningAvailabilityCost / efficiencyObservability

Real-Time Analytics & Monitoring

3 problems

Metrics Collection & Monitoring System — L7 Reference Design

Collect fleet-wide metrics, retain history, and trigger alerts quickly without drowning the system in high-cardinality series.

StreamingObservabilityPartitioningAvailabilityCost / efficiency

Ad Click / Hashtag Aggregator — L7 System Design

Ingest a high-volume event stream, serve realtime dashboards, and still produce audit-grade historical counts.

StreamingPartitioningConsistencyObservabilityCost / efficiency

Design a CI/CD Pipeline with Smart Test-Failure Handling

Run monorepo builds and tests efficiently while distinguishing genuine failures from flaky or infra-driven noise.

SchedulingObservabilityAvailabilityCost / efficiency

Messaging & Notifications

2 problems

Global Real-Time Notification System — L7 System Design

Deliver push, email, and SMS to a global user base with preferences, fallback policies, and provider-aware retries.

Fanout / deliveryAvailabilityConsistencyObservabilityCost / efficiency

Design an Online Chat / Messaging System

Support one-to-one and group chat with ordering, presence, offline sync, and read receipts.

Fanout / deliveryStateful hot pathAvailabilityStreamingConsistency

Orchestration & Coordination

1 problem

Q10

Distributed Job Scheduler (Airflow-style)

Execute DAGs on schedule, recover from worker failures, and scale to a very large number of task runs per day.

SchedulingConsistencyAvailabilityObservabilityPartitioning

Product & Edge Systems

4 problems

Q11

Design a Menu Update System for a Global Restaurant Chain

Push daily menu updates from HQ to thousands of restaurant locations and keep every device in a store on the same version.

Edge / offlineConsistencyAvailabilitySafety / correctnessObservability

Q12

Design Google Street View Image Ingestion

Ingest imagery from vehicle fleets over unreliable networks and feed geo-indexed processing pipelines downstream.

Edge / offlineStreamingPartitioningAvailabilityCost / efficiency

Q13

Mall Entrance Occupancy Management System

Keep a mall below legal occupancy limits in real time across many gates without ever blocking exits.

Safety / correctnessConsistencyEdge / offlineAvailabilityObservability

Q14

Design a Video Streaming Service

Support creator uploads, transcoding, adaptive-bitrate playback, CDN delivery, and watch-history-driven recommendations.

Fanout / deliveryStreamingCost / efficiencyAvailabilityPartitioning