System Design
Most system design material is either a glossary you cannot learn from or a case study that assumes you already know the vocabulary. This is the whole path in order: the words first, then the arithmetic, then the moving parts, then twenty-nine full designs that use all of it. Work through it one topic at a time and tick each one off as you go.
3 of 376 topics have their article written so far. The rest are on the list, in this order.
Part 10/20
Introduction to System Design
The vocabulary. Every later part assumes you can say what availability, durability and latency actually mean.
- 4System Design vs Software Architecture
- 5Functional Requirements
- 6Non-Functional Requirements
- 7Constraints and Assumptions
- 8Scalability
- 9Performance
- 10Latency
- 11Throughput
- 12Availability
- 13Reliability
- 14Durability
- 15Fault Tolerance
- 16Maintainability
- 17Extensibility
- 18Security
- 19Cost and Operational Complexity
- 20Understanding Service-Level Indicators, Objectives and Agreements
Part 20/15
System Design Interview Process
How the 45 minutes actually run, and what the person across the table is scoring.
- 21How System Design Interviews Work
- 22What Interviewers Evaluate
- 23How to Clarify an Ambiguous Problem
- 24Defining the Scope
- 25Identifying Core Use Cases
- 26Gathering Scale Requirements
- 27Making and Communicating Assumptions
- 28Designing the Simplest Initial Architecture
- 29Finding Bottlenecks
- 30Discussing Trade-Offs
- 31Handling Interviewer Requirement Changes
- 32Managing a 45-Minute Interview
- 33How to Communicate While Whiteboarding
- 34How to Finish and Summarise a Design
- 35Common System Design Interview Mistakes
Part 30/12
Back-of-the-Envelope Estimation
Arithmetic you can do out loud without a calculator, and the confidence to round hard.
- 36Why Capacity Estimation Matters
- 37Powers of Ten and Common Data Sizes
- 38Estimating Daily and Monthly Active Users
- 39Calculating Average and Peak QPS
- 40Estimating Read and Write Traffic
- 41Estimating Concurrent Connections
- 42Estimating Storage Requirements
- 43Estimating Memory and Cache Size
- 44Estimating Network Bandwidth
- 45Estimating Future Growth
- 46Avoiding False Precision in Interviews
- 47Complete Capacity Estimation Example
Part 40/26
Networking and Communication
What happens between the client pressing a key and your server seeing bytes.
- 48How a Request Travels Through a System
- 49Domain Name System
- 50IP Addresses, Ports and Sockets
- 51TCP vs UDP
- 52HTTP and HTTPS
- 53HTTP/1.1 vs HTTP/2 vs HTTP/3
- 54TLS and SSL Termination
- 55Forward Proxy vs Reverse Proxy
- 56Load Balancers
- 57Layer 4 vs Layer 7 Load Balancing
- 58Load-Balancing Algorithms
- 59Health Checks
- 60Connection Pooling
- 61Keep-Alive Connections
- 62REST APIs
- 63GraphQL
- 64gRPC
- 65WebSockets
- 66Server-Sent Events
- 67Long Polling
- 68Webhooks
- 69Synchronous vs Asynchronous Communication
- 70Service Discovery
- 71API Gateways
- 72API Versioning
- 73Pagination: Offset vs Cursor
Part 50/38
Data Storage and Databases
The largest part, because the database is the decision you cannot walk back.
- 74How to Choose a Database
- 75Relational Databases
- 76NoSQL Databases
- 77SQL vs NoSQL
- 78Key-Value Databases
- 79Document Databases
- 80Wide-Column Databases
- 81Graph Databases
- 82Time-Series Databases
- 83Search Databases
- 84Object Storage
- 85File Storage vs Block Storage vs Object Storage
- 86Database Schema Design
- 87Normalisation
- 88Denormalisation
- 89Primary and Foreign Keys
- 90Database Indexes
- 91Composite and Covering Indexes
- 92B-Trees and LSM Trees
- 93Database Transactions
- 94ACID Properties
- 95Database Isolation Levels
- 96Dirty Reads, Non-Repeatable Reads and Phantom Reads
- 97Optimistic vs Pessimistic Locking
- 98Database Replication
- 99Leader-Follower Replication
- 100Multi-Leader and Leaderless Replication
- 101Read Replicas
- 102Database Partitioning
- 103Horizontal vs Vertical Partitioning
- 104Database Sharding
- 105Choosing a Shard Key
- 106Hot Partitions
- 107Rebalancing Shards
- 108Multi-Tenant Database Design
- 109Database Connection Pooling
- 110Database Migrations at Scale
- 111Soft Deletion, Archiving and Data Retention
Part 60/18
Caching
The cheapest performance win and the richest source of subtle bugs.
- 112What Is Caching?
- 113Why Caching Improves Performance
- 114Browser, CDN, Application and Database Caching
- 115Cache-Aside Pattern
- 116Read-Through Cache
- 117Write-Through Cache
- 118Write-Behind Cache
- 119Cache Invalidation
- 120Time to Live
- 121Cache Eviction Policies
- 122Cache Consistency
- 123Cache Stampede
- 124Cache Penetration
- 125Cache Avalanche
- 126Hot Keys
- 127Distributed Caching
- 128Redis in System Design
- 129When Caching Makes a System Worse
Part 70/23
Message Queues and Event-Driven Systems
Decoupling work from the request that asked for it, and the delivery guarantees that follow.
- 130What Is a Message Queue?
- 131Why We Use Asynchronous Processing
- 132Message Queues vs Event Streams
- 133Producers, Consumers and Brokers
- 134Consumer Groups
- 135Message Ordering
- 136At-Most-Once Delivery
- 137At-Least-Once Delivery
- 138Exactly-Once and Effectively-Once Processing
- 139Idempotency
- 140Retries and Exponential Backoff
- 141Dead-Letter Queues
- 142Backpressure
- 143Kafka vs RabbitMQ vs Cloud Queues
- 144Publish-Subscribe Systems
- 145Event-Driven Architecture
- 146Event Sourcing
- 147CQRS
- 148Saga Pattern
- 149Transactional Outbox Pattern
- 150Change Data Capture
- 151Stream Processing
- 152Batch Processing vs Stream Processing
Part 80/22
Distributed Systems Fundamentals
The theory that explains why the obvious design does not work across a network.
- 153What Is a Distributed System?
- 154Challenges of Distributed Systems
- 155Network Partitions
- 156Strong Consistency
- 157Eventual Consistency
- 158Read-After-Write and Causal Consistency
- 159CAP Theorem
- 160PACELC Theorem
- 161Quorum Reads and Writes
- 162Consensus
- 163Leader Election
- 164Distributed Locks
- 165Consistent Hashing
- 166Clock and Time Problems
- 167Logical and Vector Clocks
- 168Conflict Detection and Resolution
- 169Split-Brain Problems
- 170Gossip Protocol
- 171Data Locality
- 172Distributed Transactions
- 173Two-Phase Commit
- 174Handling Partial Failure
Part 90/22
Architecture and Scaling
Shapes a system can take, and what each shape costs to run.
- 175Vertical vs Horizontal Scaling
- 176Stateless vs Stateful Services
- 177Monolithic Architecture
- 178Modular Monolith
- 179Microservices Architecture
- 180Monolith vs Microservices
- 181Service Boundaries
- 182Domain-Driven Design
- 183Service-Oriented Architecture
- 184Serverless Architecture
- 185Containers
- 186Container Orchestration
- 187Autoscaling
- 188Session Management in Distributed Systems
- 189Sticky Sessions
- 190Multi-Region Architecture
- 191Active-Active vs Active-Passive Architecture
- 192Edge Computing
- 193Cell-Based Architecture
- 194Control Plane vs Data Plane
- 195Multi-Tenant Architecture
- 196Build vs Buy Decisions
Part 100/20
Reliability and Resilience
Assuming every dependency fails, and designing so the failure stays small.
- 197Designing for Failure
- 198Single Points of Failure
- 199Timeouts
- 200Retries
- 201Circuit Breakers
- 202Bulkhead Pattern
- 203Graceful Degradation
- 204Load Shedding
- 205Rate Limiting
- 206Throttling
- 207Redundancy
- 208Failover
- 209Disaster Recovery
- 210Recovery Point Objective and Recovery Time Objective
- 211Backup and Restore Strategies
- 212Idempotency Keys
- 213Handling Duplicate Requests
- 214Preventing Cascading Failures
- 215Chaos Engineering
- 216Zero-Downtime Architecture
Part 110/22
Common Specialised Systems
Named building blocks that recur across interview problems.
- 217Content Delivery Networks
- 218Full-Text Search
- 219Inverted Indexes
- 220Search Ranking
- 221Autocomplete Systems
- 222Recommendation Systems
- 223News Feed Generation
- 224Fan-Out on Write vs Fan-Out on Read
- 225Real-Time Chat Systems
- 226Presence and Online Status
- 227Notification Systems
- 228Location-Based Services
- 229Geohashing
- 230Payment Systems
- 231Inventory and Reservation Systems
- 232Distributed Job Schedulers
- 233Web Crawlers
- 234Media Upload and Processing
- 235Video Streaming
- 236Collaborative Editing
- 237Analytics and Metrics Pipelines
- 238Logging Platforms
Part 120/19
Security
The part candidates skip and interviewers remember.
- 239Authentication vs Authorisation
- 240Sessions vs Tokens
- 241JSON Web Tokens
- 242OAuth 2.0 and OpenID Connect
- 243Role-Based and Attribute-Based Access Control
- 244Encryption in Transit and at Rest
- 245Secrets Management
- 246API Security
- 247Rate Limiting and Abuse Prevention
- 248DDoS Protection
- 249Secure File Uploads
- 250Signed URLs
- 251Webhook Security
- 252Replay-Attack Prevention
- 253Tenant Isolation
- 254Personally Identifiable Information
- 255Data Retention and Deletion
- 256Audit Logging
- 257Threat Modelling
Part 130/20
Observability and Operations
Knowing what the system is doing, and changing it without downtime.
- 258Logs, Metrics and Traces
- 259Structured Logging
- 260Correlation IDs
- 261Distributed Tracing
- 262Important System Metrics
- 263SLIs, SLOs and Error Budgets
- 264Alerting
- 265Dashboards
- 266Health, Readiness and Liveness Checks
- 267Capacity Planning
- 268Performance Testing
- 269Load, Stress, Spike and Soak Testing
- 270Profiling and Bottleneck Analysis
- 271Feature Flags
- 272Rolling Deployments
- 273Blue-Green Deployments
- 274Canary Deployments
- 275Rollbacks
- 276Infrastructure as Code
- 277Cost Optimisation
Part 140/30
Applied AI System Design
What changes when a model is in the request path: cost, latency, and non-determinism.
- 278How AI System Design Differs from Traditional System Design
- 279Online vs Offline Inference
- 280Model-Serving Architecture
- 281CPU vs GPU Inference
- 282Model Batching
- 283Model Gateways and Routers
- 284Model and Provider Fallback
- 285Prompt and Model Versioning
- 286Streaming AI Responses
- 287Semantic Caching
- 288Embedding Pipelines
- 289Vector Databases in System Design
- 290RAG System Architecture
- 291Document Ingestion Pipelines
- 292Chunking and Document Versioning
- 293Hybrid Search and Reranking
- 294Agent Architecture
- 295Agent Orchestration
- 296Tool Execution and Permissions
- 297Workflow State and Checkpointing
- 298Human-in-the-Loop Systems
- 299AI Guardrails
- 300AI Evaluation Systems
- 301Hallucination and Faithfulness Evaluation
- 302Abstention and Confidence Handling
- 303AI Observability
- 304AI Cost and Latency Optimisation
- 305Multi-Tenant AI Platforms
- 306Data Privacy in AI Systems
- 307Designing Reliable Agent Workflows
Part 150/19
Low-Level Design
Class-level design: the other interview, and the one that rewards precision.
- 308What Is Low-Level Design?
- 309Object-Oriented Design
- 310SOLID Principles
- 311Composition vs Inheritance
- 312Interfaces and Abstractions
- 313Dependency Injection
- 314Common Design Patterns
- 315Factory Pattern
- 316Strategy Pattern
- 317Observer Pattern
- 318Adapter Pattern
- 319Decorator Pattern
- 320State Pattern
- 321Repository Pattern
- 322State Machines
- 323Domain Modelling
- 324Concurrency and Thread Safety
- 325Extensible API and Class Design
- 326Testing a Low-Level Design
Part 160/29
System Design Case Studies
Full problems, worked end to end. This is where the earlier parts get used.
- 327Design a URL Shortener
- 328Design a Rate Limiter
- 329Design Pastebin
- 330Design a Notification System
- 331Design a Distributed Job Scheduler
- 332Design a Web Crawler
- 333Design Cloud File Storage
- 334Design a Logging System
- 335Design a Metrics Platform
- 336Design a Social Media Feed
- 337Design a Chat Application
- 338Design a Video-Streaming Platform
- 339Design a Ride-Sharing Platform
- 340Design an E-Commerce Platform
- 341Design a Payment Platform
- 342Design a Ticket-Booking System
- 343Design a Collaborative Document Editor
- 344Design a Search Engine
- 345Design an Autocomplete Service
- 346Design a Multi-Tenant SaaS Platform
- 347Design an API Gateway
- 348Design a Distributed Cache
- 349Design a Message Broker
- 350Design a ChatGPT-Like Application
- 351Design a Multi-Tenant RAG Platform
- 352Design an AI Coding Assistant
- 353Design an AI Document-Processing Platform
- 354Design an Agent-Workflow Platform
- 355Design an AI Evaluation Platform
Part 170/10
Low-Level Design Exercises
The classic object-modelling problems, each small enough to finish in one sitting.
- 356Design a Parking Lot
- 357Design an Elevator System
- 358Design a Vending Machine
- 359Design a Library System
- 360Design a Logger
- 361Design a Task Scheduler
- 362Design a Notification Framework
- 363Design a Payment State Machine
- 364Design a Ride-Booking System
- 365Design a Board Game
Part 180/11
Final Interview Preparation
Rehearsal. A repeatable template and the drills that make it automatic.
- 366Creating a Reusable Whiteboard Template
- 367Practising Requirement Clarification
- 368Practising Capacity Estimation
- 369Practising APIs and Data Models
- 370Practising Critical Request Flows
- 371Practising Failure Scenarios
- 372Practising Trade-Off Explanations
- 373Conducting Timed Mock Interviews
- 374Handling Follow-Up Questions
- 375Evaluating Your Own Performance
- 376Final System Design Interview Checklist