Fundamentals

Master the building blocks that appear in every system design interview.

---

Why Fundamentals Matter

Before designing complex systems like Twitter or Uber, you need to understand the components that make them work. These fundamentals are the "vocabulary" of system design:

• Interview Framework teaches you HOW to approach any system design question
• Estimation sizes infrastructure before building
• Networking enables communication between services
• Databases persist your data reliably and at scale
• Caches store frequently accessed data for lightning-fast retrieval
• Load Balancers distribute traffic so no single server gets overwhelmed
• API Design defines how services communicate clearly and reliably
• Concurrency handles multiple operations simultaneously
• Security protects data and systems from threats
• Scalability & Reliability ensures systems handle growth and failures
• Distributed Systems coordinate work across multiple machines

---

What's Covered

Topic	Description	Difficulty
Interview Framework	How to approach any system design question	⭐ Beginner
Estimation & Planning	Back-of-the-envelope calculations	⭐⭐ Intermediate
Networking	TCP/UDP, HTTP, DNS, WebSockets	⭐⭐ Intermediate
Databases	SQL vs NoSQL, ACID, CAP, scaling	⭐⭐ Intermediate
Caching	Store data for faster access	⭐ Beginner
Load Balancing	Distribute requests across multiple servers	⭐ Beginner
API Design	REST, GraphQL, versioning, rate limiting	⭐⭐ Intermediate
Concurrency	Threads, locks, async patterns	⭐⭐ Intermediate
Security	Encryption, hashing, TLS, vulnerabilities	⭐⭐ Intermediate
Scalability & Reliability	Scaling, availability, disaster recovery	⭐⭐⭐ Advanced
Distributed Systems	CAP, consensus, DHTs, message queues	⭐⭐⭐ Advanced

---

How to Study These

1. Read through each topic - Understand the concepts, not just the terms
2. Know the trade-offs - Every choice has pros and cons
3. Be ready to apply them - Interviewers will ask "which would you use here?"

Tip

When studying, ask yourself: "When would I use this? When would I NOT use this?"

---

Quick Reference Card

Print this out or keep it handy:

LOAD BALANCING
├── Round Robin      → Simple, equal distribution
├── Least Connections → Best for varying request durations
├── IP Hash          → Session persistence (sticky sessions)
└── Health Checks    → Remove unhealthy servers automatically

CACHING
├── Cache-Aside      → App checks cache, then DB
├── Write-Through    → Write to cache AND DB together
├── Write-Behind     → Write cache, async to DB
└── Eviction: LRU, LFU, TTL

DATABASES
├── SQL              → ACID, relationships, complex queries
├── NoSQL            → Flexibility, horizontal scaling
├── Sharding         → Split data across multiple DBs
└── Replication      → Copy data for availability

NETWORKING
├── TCP              → Reliable, ordered (HTTP, databases)
├── UDP              → Fast, unreliable (video, gaming)
├── HTTP/REST        → Request-response, stateless
├── gRPC             → High-performance microservices
└── WebSocket        → Bidirectional, real-time

CONCURRENCY
├── Threads          → Shared memory, fast communication
├── Processes        → Isolated memory, crash safety
├── Mutex/Lock       → Exclusive access to resource
├── Thread Pool      → Reuse threads, bound resources
└── Async I/O        → Single thread, many I/O operations

DISTRIBUTED SYSTEMS
├── CAP Theorem      → Consistency vs Availability vs Partition Tolerance
├── Consensus        → Raft (understandable), Paxos (foundational)
├── Consistent Hash  → Minimal key remapping on node changes
└── Message Queues   → Kafka (log), RabbitMQ (broker), SQS (managed)

API DESIGN
├── REST             → Resource-oriented, HTTP semantics
├── GraphQL          → Client-specified queries, single endpoint
├── Versioning       → URI path (v1/v2), header, query param
└── Auth             → JWT (stateless), OAuth (delegated), API keys

SECURITY
├── Encryption       → AES (symmetric), RSA/ECDSA (asymmetric)
├── Hashing          → bcrypt/Argon2 (passwords), SHA-256 (integrity)
├── TLS              → Encrypt in transit, certificate chain
└── Input Validation → Parameterized queries, sanitization

SCALABILITY & RELIABILITY
├── Horizontal Scale → Stateless services + load balancer
├── Redundancy       → Active-active, active-passive
├── Circuit Breaker  → Stop cascading failures
└── Monitoring       → Latency, traffic, errors, saturation

ESTIMATION
├── Traffic          → DAU × actions/user ÷ 86,400 = QPS
├── Storage          → writes/day × size × 365 × replication
├── Bandwidth        → QPS × response_size
└── Rule of thumb    → 1M req/day ≈ 12 QPS

---

Interview Cheat Sheet

When They Ask About...	Think About...
Handling traffic	Load balancing, horizontal scaling, auto-scaling
Making it fast	Caching, CDN, read replicas
Storing data	SQL vs NoSQL, consistency needs, sharding
Real-time features	WebSockets, message queues, pub/sub
Handling failures	Replication, circuit breakers, retries, failover
Concurrent access	Locks, optimistic concurrency, idempotency
Multiple services	Consensus protocols, leader election, DHTs
API design	REST vs GraphQL vs gRPC, versioning, pagination
Protecting data	TLS, encryption at rest, input validation, auth
How big should it be?	Back-of-envelope estimation, capacity planning

---

What's Next?

After mastering these fundamentals:

1. Go deeper with Advanced Topics for Senior/Staff-level concepts (13 topics)
2. Apply them in System Design Examples — 24 walkthroughs organized by learning path
3. Learn ML/GenAI in GenAI/ML Fundamentals (7 topics) → ML System Design (10 designs) → GenAI System Design (10 designs with interview transcripts)
4. Practice drawing architectures using these building blocks