System Design · Deep Dives

Design systems beyond the whiteboard.

Interview-depth case studies where every decision is justified. Built as a progression — start from first principles, climb to distributed systems.

Concepts
Back of Envelope Estimation
Apply the Concepts — Case Studies
01
Foundation ● ○ ○ ○ ○

First principles of scale

Unique ID schemes, sharding strategies, read-heavy caching, hot key mitigation, pre-generated key pools, tiered storage.

Snowflake ID UUID Clock skew Consistent hashing Bloom filters KGS Hot key problem Cold storage
Start
02
Ascent ● ● ○ ○ ○

Async pipelines & multi-channel delivery

Kafka fan-out, per-channel workers, retry/DLQ, scheduled dispatch, content deduplication, async uploads, expiry at scale.

Kafka fan-out Retry / DLQ Bloom filter Circuit breaker Content-addressable storage Async S3 upload TTL sweeper Watchdog
Open
03
Expedition ● ● ● ○ ○

Distributed rate control

Rate limiting algorithms from scratch, atomic Redis counters, distributed enforcement, rule management, DDoS layered defence.

Token bucket Leaky bucket Sliding window Fixed window Lua atomicity Redis cluster DDoS defence
Open
04
Summit ● ● ● ● ○

Storage engine internals

Leaderless replication, quorum reads/writes, LSM tree internals, compaction strategies, gossip membership, conflict resolution.

LSM trees Compaction Bloom filters Quorum W/R Gossip protocol Hinted handoff Vector clocks Anti-entropy
Open
05
Battleground ● ● ● ● ●

Real-time systems at scale

WebSocket connection management, per-conversation sequencing, offline queues, thundering herd, backpressure, hot partition salting.

WebSockets Per-conv. sequencing Offline delivery Thundering herd Backpressure Hot partition salting Reconnect storm Read receipts
Open