Patternwise

HLD track

System Design

Design systems that scale, survive failure, and hold up under an interviewer's follow-ups.

A path from networking fundamentals to multi-region architecture, written for engineers aiming at senior system-design rounds. Each lesson builds the vocabulary and trade-off instincts you need to drive a 45-minute design discussion with confidence.

4stages
16modules
46lessons
22problems
7hours
Start the first lesson
0/46 lessons
Stage 1

Foundations

How requests actually travel, how clients talk to servers, and how to put numbers on a design before you draw a single box.

Module 013 lessons · 26 min

How the Web Works

Trace a request from a browser to a server and explain where DNS, TCP, TLS, and HTTP each add latency.

  1. ✓DNS: The Internet's Phone BookHow a hostname becomes an IP address, and how DNS doubles as a traffic-steering tool.8m
  2. ✓HTTP, HTTPS, and TLSRequest/response semantics, HTTP/1.1 vs 2 vs 3, and what the TLS handshake costs you.10m
  3. ✓TCP vs UDPReliable ordered streams versus fire-and-forget datagrams, and when each one wins.8m
Module 023 lessons · 28 min

Client-Server Communication & APIs

Choose between REST, gRPC, GraphQL, and push-style protocols, and justify the choice for a given client.

  1. ✓REST API DesignResources, verbs, status codes, pagination, and versioning for APIs that age well.10m
  2. ✓gRPC and GraphQLBinary RPC for service-to-service calls and flexible queries for client-facing aggregation.9m
  3. ✓WebSockets, SSE, and Long PollingThree ways to push data to clients and the operational cost of each.9m
Module 033 lessons · 30 min

Estimation, Performance & Availability

Do quick capacity math, reason about latency and concurrency with Little's law, and translate nines into downtime budgets.

  1. ✓Back-of-the-Envelope EstimationTurn user counts into QPS, storage, and bandwidth with the numbers every engineer should know.12m
  2. ✓Latency, Throughput & Little's LawWhy percentiles matter and how concurrency, rate, and wait time are tied together.9m
  3. ✓Availability & SLA MathWhat each extra nine costs, and how availability composes in series and in parallel.9m
Stage 2

Core System Design

The building blocks that show up in nearly every design: load balancers, caches, databases, queues, and storage.

Module 042 lessons · 17 min

Load Balancing

Place L4 and L7 load balancers correctly and pick an algorithm that matches the workload.

  1. ✓Load Balancers: L4 vs L7What a load balancer does, where it sits, and what you gain by looking inside requests.9m
  2. ✓Load Balancing AlgorithmsRound robin, least connections, power of two choices, and hashing for affinity.8m
Module 054 lessons · 36 min

Caching

Design a cache layer with the right write strategy, eviction policy, and protections against stampedes and hot keys.

  1. ✓Caching StrategiesCache-aside, read-through, write-through, write-behind, and where each one fits.10m
  2. ✓Eviction & InvalidationLRU, LFU, TTLs, and the hard problem of keeping a cache honest.9m
  3. ✓Cache Stampedes & Hot KeysWhat happens when a popular key expires or one key gets all the traffic.9m
  4. ✓CDNs & Edge CachingServing bytes close to users and controlling what the edge keeps.8m
Module 065 lessons · 50 min

Databases

Pick a data store, design indexes, and scale it with replication, sharding, and consistent hashing.

  1. ✓SQL vs NoSQLRelational, key-value, document, wide-column, and graph stores compared on real criteria.10m
  2. ✓Indexing: B-Trees vs LSM TreesHow storage engines organize data on disk and why that shapes read and write costs.11m
  3. ✓ReplicationLeader-follower, multi-leader, and leaderless replication, plus replication lag pitfalls.10m
  4. ✓Partitioning & ShardingSplitting data across machines by range or hash, and handling hot partitions.10m
  5. ✓Consistent HashingPlacing keys on a ring so adding a node moves only a small slice of data.9m
Module 072 lessons · 21 min

Message Queues & Streams

Choose between Kafka, RabbitMQ, and SQS and build consumers that are safe under redelivery.

  1. ✓Queues vs Logs: Kafka, RabbitMQ, SQSWork queues versus replayable logs, and how the three popular systems differ.11m
  2. ✓Delivery Semantics & IdempotencyAt-most-once, at-least-once, effectively-once, and how idempotent consumers close the gap.10m
Module 082 lessons · 17 min

Storage Systems

Decide between object, block, and file storage and design large-file upload paths.

  1. ✓Blob & Object StorageHow S3-style stores work and patterns like presigned URLs and multipart upload.9m
  2. ✓Block Storage & Distributed File SystemsVolumes, shared file systems, and the chunked design behind HDFS-style systems.8m
Stage 3

Production Systems

What changes when a design has to run for years: consistency trade-offs, traffic control, operability, resilience, and security.

Module 092 lessons · 19 min

Distributed Systems Trade-offs

Explain CAP and PACELC precisely and pick the consistency model a feature actually needs.

  1. ✓CAP & PACELCWhat the CAP theorem really says, and the latency trade-off PACELC adds.9m
  2. ✓Consistency ModelsLinearizable, sequential, causal, read-your-writes, and eventual consistency explained.10m
Module 102 lessons · 19 min

Traffic Management

Implement rate limiting and route traffic through gateways and service discovery.

  1. ✓Rate Limiting AlgorithmsToken bucket, leaky bucket, fixed and sliding windows compared.10m
  2. ✓API Gateways & Service DiscoveryA single front door for clients, and how services find each other behind it.9m
Module 112 lessons · 19 min

Architecture & Observability

Argue for the right service granularity and instrument a system so you can tell when it is unhealthy.

  1. ✓Microservices vs MonolithThe real costs of splitting a system, and the modular monolith as a middle path.9m
  2. ✓Observability & SLOsLogs, metrics, and traces, plus SLIs, SLOs, and error budgets.10m
Module 123 lessons · 29 min

Resilience & Security

Contain failures with timeouts, retries, circuit breakers, and bulkheads, and secure APIs with OAuth2 and JWTs.

  1. ✓Timeouts, Retries, Backoff & JitterRetrying without turning a blip into an outage.9m
  2. ✓Circuit Breakers & BulkheadsFailing fast and isolating resources so one bad dependency cannot sink everything.9m
  3. ✓AuthN, AuthZ, OAuth2 & JWTIdentity versus permission, delegated access, and the trade-offs of stateless tokens.11m
Stage 4

Advanced System Design

Coordination, specialized data systems, global deployments, and a repeatable framework for the interview itself.

Module 135 lessons · 48 min

Transactions, Consensus & IDs

Keep data consistent across services, explain how Raft elects leaders, and generate unique IDs at scale.

  1. ✓Distributed Transactions & 2PCAtomic commit across nodes, and why two-phase commit is rarely used across services.9m
  2. ✓Sagas & the Transactional OutboxLong-running workflows with compensations, and publishing events without dual writes.11m
  3. ✓Consensus & RaftHow a cluster agrees on an ordered log even when nodes crash.11m
  4. ✓Leader Election & CoordinationLeases, fencing tokens, and coordination services like ZooKeeper and etcd.9m
  5. ✓Distributed ID GenerationUUIDs, ticket servers, and Snowflake-style time-ordered IDs.8m
Module 143 lessons · 30 min

Search, Real-time & Geo

Design search indexes, real-time fan-out with presence, and location queries with geohashes or quadtrees.

  1. ✓Search & Inverted IndexesTokenizing, posting lists, ranking, and sharding a search cluster.10m
  2. ✓Real-time Fan-out, Pub/Sub & PresenceDelivering events to millions of connected clients and tracking who is online.10m
  3. ✓Geohash & QuadtreesTurning 2D location queries into something a normal index can answer.10m
Module 153 lessons · 29 min

Data Pipelines & Global Scale

Build batch and streaming pipelines fed by CDC, and design for regional failure with clear RPO/RTO targets.

  1. ✓Batch vs Stream ProcessingBounded versus unbounded data, windows, watermarks, and late events.10m
  2. ✓Change Data CaptureStreaming a database's changes to caches, search, and warehouses.8m
  3. ✓Multi-Region & Disaster RecoveryActive-passive versus active-active, RPO and RTO, and data residency.11m
Module 162 lessons · 21 min

The System Design Interview

Run a 45-minute design round end to end, with clear time boxes and well-argued trade-offs.

  1. ✓A Framework for the 45-Minute RoundRequirements, estimates, API, data model, high-level design, deep dives, and wrap-up.12m
  2. ✓Driving Deep Dives & Trade-offsChoosing what to go deep on and how to talk about trade-offs like a senior engineer.9m

Then practise

Design Problems

Classic system-design questions worked end to end, from requirements and estimates to deep dives. Try each one on a whiteboard first, then compare your design against the walkthrough.

Open design problems →