System Design Q&A — Question Bank

Format: Every answer is a chain of sequential whys — from ground up, no gaps, prose over bullets. One file per question. Answers added one by one.

Part 1: Networking & Communication

Q01 — Why does TCP exist — what problem does it solve that the raw internet doesn't? ✅
Q02 — Why does UDP exist — when is "unreliable" actually the right choice? ✅
Q03 — How did HTTP evolve from 1.1 → HTTP/2 → HTTP/3 — what problem does each version fix? ✅
Q04 — Why do WebSockets exist — what does HTTP get wrong for real-time communication? ✅
Q05 — What is long polling, and why did it exist before WebSockets? ✅
Q06 — What is Server-Sent Events (SSE) and how does it differ from WebSockets? ✅
Q07 — How does DNS work — what actually happens between typing a URL and a response arriving? ✅
Q08 — What is a load balancer — why can't you just add more servers without one? ✅

Q09 — Why do databases exist — why not just use files? ✅
Q10 — How do database indexes actually work — what is a B-tree doing physically? ✅
Q11 — Why does a full table scan sometimes beat an index? ✅
Q12 — What is an LSM tree — why do some databases choose it over a B-tree? ✅
Q13 — Why do databases use write-ahead logs (WAL)? ✅
Q14 — What does ACID really mean — not the acronym, but the actual guarantees and trade-offs? ✅
Q15 — How does a database transaction work under the hood? ✅
Q16 — What is connection pooling and why does it exist? ✅

Q17 — Why is caching hard — it's not just "store it in memory"? ✅
Q18 — What is cache-aside, write-through, and write-behind — why does the choice matter? ✅
Q19 — What is a cache stampede — and how do you prevent it? ✅
Q20 — How do cache eviction policies work — LRU, LFU — and why does the choice matter? ✅
Q21 — Why do CDNs exist — and what can they not solve? ✅

Q27 — What is optimistic locking vs pessimistic locking — and when does each break? ✅
Q28 — Why is distributed locking hard, and how do you do it safely? ✅
Q29 — How do you design a retry system that doesn't destroy your own service? ✅
Q30 — What is a circuit breaker and why does it exist? ✅
Q31 — How does rate limiting work — and what are the failure modes of each algorithm? ✅
Q32 — What is backpressure and why does ignoring it kill systems under load? ✅
Q33 — What is the difference between a timeout, a deadline, and a retry budget? ✅

Q34 — What is consistent hashing — why does naive hashing break in distributed systems? ✅
Q35 — How does replication work — and what is replication lag? ✅
Q36 — Why do we shard databases — and what breaks when we do? ✅
Q37 — What is eventual consistency — why does it exist and what does it cost you? ✅
Q38 — What is strong consistency — and what does it cost you? ✅
Q39 — What is the CAP theorem — and why is it almost always misunderstood? ✅
Q40 — What is a distributed transaction — and why is two-phase commit (2PC) dangerous? ✅
Q41 — How do distributed systems reach consensus — what is Raft actually solving? ✅

Q42 — Why does a message queue exist — what's wrong with direct service-to-service calls? ✅
Q43 — What is the difference between at-most-once, at-least-once, and exactly-once delivery? ✅
Q44 — How does Kafka work internally — why is it fast? ✅
Q45 — What are consumer groups in Kafka — why do they exist? ✅
Q46 — What is the difference between a queue and a pub/sub system? ✅

Q47 — Why do microservices exist — and when are they the wrong choice? ✅
Q48 — What is service discovery — why can't services just hardcode each other's addresses? ✅
Q49 — What is the outbox pattern — why does it exist? ✅
Q50 — What is the saga pattern — how do you handle failures across services without 2PC? ✅
Q51 — What is event sourcing — what do you gain and what does it cost? ✅
Q52 — What is CQRS — why would you separate reads from writes at the architecture level? ✅

Q53 — What is structured logging — why does log format matter at scale? ✅
Q54 — What is the difference between metrics, logs, and traces — when does each fail you? ✅
Q55 — What is distributed tracing — why is a single log line not enough in microservices? ✅

Q56 — Why is storing passwords as hashes not enough — what is a rainbow table attack? ✅
Q57 — How does JWT work — and what are its real security pitfalls? ✅
Q58 — How does OAuth 2.0 work — why does it exist instead of just passing passwords? ✅

Q59 — How does a search index work — why can't you use a database LIKE query at scale? ✅
Q60 — How does a time-series database differ from a relational one — and why? ✅
Q61 — How would you design a distributed ID generator (like Twitter's Snowflake)? ✅