[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-77419":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":15,"stars7d":15,"stars30d":16,"stars90d":15,"forks30d":15,"starsTrendScore":15,"compositeScore":17,"rankGlobal":10,"rankLanguage":10,"license":18,"archived":19,"fork":19,"defaultBranch":20,"hasWiki":21,"hasPages":19,"topics":22,"createdAt":10,"pushedAt":10,"updatedAt":23,"readmeContent":24,"aiSummary":25,"trendingCount":15,"starSnapshotCount":15,"syncStatus":26,"lastSyncTime":27,"discoverSource":28},77419,"hld-handbook","handbook-academy\u002Fhld-handbook","handbook-academy","Open-source High-Level Design & System Design handbook. 181 chapters, 727K words, 653 diagrams, CC BY-SA 4.0.","https:\u002F\u002Fhld.handbook.academy",null,"JavaScript",183,5,59,0,122,2.33,"Creative Commons Attribution Share Alike 4.0 International",false,"main",true,[],"2026-06-12 02:03:43","\u003Cdiv align=\"center\">\n\n\u003Cimg src=\"logo.svg\" alt=\"The HLD Handbook\" width=\"160\" \u002F>\n\n# The HLD Handbook\n\n**The open-source High-Level Design & System Design handbook.**\n\n159 chapters + 22 trade-off pages · 181 pages · 772,000 words · 719 diagrams · 3,100+ citations\n\n[![License: CC BY-SA 4.0](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-CC%20BY--SA%204.0-blue.svg)](LICENSE)\n[![PRs Welcome](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPRs-welcome-brightgreen.svg)](CONTRIBUTING.md)\n[![CI](https:\u002F\u002Fgithub.com\u002Fhandbook-academy\u002Fhld-handbook\u002Factions\u002Fworkflows\u002Fcontent-ci.yml\u002Fbadge.svg)](https:\u002F\u002Fgithub.com\u002Fhandbook-academy\u002Fhld-handbook\u002Factions)\n[![GitHub stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fhandbook-academy\u002Fhld-handbook?style=social)](https:\u002F\u002Fgithub.com\u002Fhandbook-academy\u002Fhld-handbook\u002Fstargazers)\n\n**Read free at [hld.handbook.academy](https:\u002F\u002Fhld.handbook.academy)** *(public beta)*\n\n[Start reading](#start-reading) · [Full curriculum](#the-full-curriculum) · [Case studies](#part-8--case-studies-56-chapters) · [Trade-offs](#trade-offs-library-22-pages) · [Contributing](#contributing)\n\n\u003C\u002Fdiv>\n\n---\n\n## Table of Contents\n\n- [What this is](#what-this-is)\n- [Why this exists](#why-this-exists)\n- [Who this is for](#who-this-is-for)\n- [How the HLD Handbook compares](#how-the-hld-handbook-compares)\n- [Start reading](#start-reading)\n- [The full curriculum](#the-full-curriculum)\n  - [Part 0 — Prerequisites (5 chapters)](#part-0--prerequisites-5-chapters)\n  - [Part 1 — Core Fundamentals (7 chapters)](#part-1--core-fundamentals-7-chapters)\n  - [Part 2 — Building Blocks (16 chapters)](#part-2--building-blocks-16-chapters)\n  - [Part 3 — Distributed Systems Theory (11 chapters)](#part-3--distributed-systems-theory-11-chapters)\n  - [Part 4 — Data Systems (10 chapters)](#part-4--data-systems-10-chapters)\n  - [Part 5 — Architecture Patterns (11 chapters)](#part-5--architecture-patterns-11-chapters)\n  - [Part 6 — Reliability & Operations (11 chapters)](#part-6--reliability--operations-11-chapters)\n  - [Part 7 — Security at Scale (10 chapters)](#part-7--security-at-scale-10-chapters)\n  - [Part 8 — Case Studies (56 chapters)](#part-8--case-studies-56-chapters)\n  - [Part 9 — AI & ML System Design (15 chapters)](#part-9--ai--ml-system-design-15-chapters)\n  - [Part 10 — Emerging Patterns (1 chapter)](#part-10--emerging-patterns-1-chapter)\n  - [Part 11 — Interview Framework (6 chapters)](#part-11--interview-framework-6-chapters)\n  - [Trade-offs Library (22 pages)](#trade-offs-library-22-pages)\n- [Study plans](#study-plans)\n- [Project statistics](#project-statistics)\n- [Quality standards](#quality-standards)\n- [Contributing](#contributing)\n- [Project structure](#project-structure)\n- [Development setup](#development-setup)\n- [Governance](#governance)\n- [FAQ](#faq)\n- [License](#license)\n- [Citation](#citation)\n- [Acknowledgments](#acknowledgments)\n- [Community](#community)\n\n---\n\n## What this is\n\nThe **HLD Handbook** is an open-source, opinionated, end-to-end textbook on high-level software design, distributed systems, and modern infrastructure — written for engineers who need more than a cheat sheet but less than a graduate-level textbook.\n\nEvery concept in this repository is taught **inline** as a full-length article: introduction, first-principles explanation, diagrams, worked examples, trade-offs, production gotchas, and citations to primary sources. No stubs. No \"coming soon.\" No external blog redirects. No summary bullets that tell you to go somewhere else to actually learn.\n\nAt roughly **772,000 words across 181 pages**, the handbook is equivalent in scope to a **~2,200-page book** — longer than *Designing Data-Intensive Applications* and Alex Xu's *System Design Interview* volumes 1 and 2 combined. It covers everything from the TCP\u002FIP stack to LLM serving architectures, and it treats each topic at roughly the same depth so you can trust the level of detail regardless of which chapter you land on.\n\n**Read it free at [hld.handbook.academy](https:\u002F\u002Fhld.handbook.academy).** Every `.md` file under `content\u002F` also renders natively on GitHub — all 719 Mermaid diagrams display in the GitHub UI, all footnotes work, all cross-references resolve — but the website gives you search, dark mode, per-chapter diagram zoom, social cards, OG images, and fast client-side navigation across the full 181-page curriculum.\n\nThe handbook is continuously updated. Every chapter's frontmatter declares `date_created` and `date_updated`, so you can see exactly how fresh each page is. We care about numbers being right today, not right in 2022.\n\n## Why this exists\n\nThe open-source system design landscape has a shape, and that shape is frustrating:\n\n- **Link dumps** — curated lists that point at dozens of scattered blog posts, papers, and talks. Great for discovery, useless as a learning path. You end up with 60 tabs open and no through-line.\n- **Teaser-and-redirect repos** — READMEs with a few hundred words on each topic that nudge you toward a paid course hosted elsewhere. The GitHub repo is the marketing page; the actual teaching is behind a $150-$400+\u002Fyear paywall.\n- **Monolithic READMEs frozen in time** — a single 5,000-line `README.md` that was great in 2021 but hasn't kept up with the modern stack. No LLM serving, no CRDTs, no post-quantum crypto, no FinOps. And nobody wants to review a PR against a 5,000-line file.\n- **Surface-level outlines** — bullet-point summaries that tell you *what* exists without explaining *why* you'd pick one approach over another. You learn the vocabulary without the judgment.\n- **Interview-only prep** — focused on passing a specific 45-minute screen, not on actually operating systems at scale.\n\nThis handbook fixes all of that:\n\n- **100% inline content.** Every chapter is a full teaching article written from scratch, living in this repo as plain Markdown. Nothing is a stub. Nothing redirects you elsewhere to learn.\n- **Progressive curriculum.** 159 teaching chapters across 12 parts, plus a 22-page Trade-offs Library, sequenced deliberately from prerequisites (Part 0) to Staff+ topics (Parts 7-9). Each chapter declares its prerequisites, learning objectives, estimated reading time, and difficulty tier.\n- **Research-backed.** Every chapter ends with a **Further Reading & References** section citing primary sources — SIGMOD and OSDI papers, RFCs, IETF drafts, engineering postmortems, official docs, canonical books. **3,100+ citations** across the handbook. Not blog posts about blog posts.\n- **Opinionated.** Every topic picks a recommended approach and explains *why*, instead of listing options without guidance. Where reasonable people disagree, the trade-offs are made explicit with a decision flowchart and \"when to pick A\" vs \"when to pick B\" tables.\n- **Modern (2025+).** LLM serving, RAG pipelines, AI agents, multi-agent orchestration, CRDTs, edge computing, FinOps, post-quantum cryptography, platform engineering, local-first software, differential privacy, MCP — all first-class topics, not afterthoughts.\n- **Decision-oriented.** The **Trade-offs Library** gives decision flowcharts for the hardest architectural choices: SQL vs NoSQL, REST vs gRPC vs GraphQL, monolith vs microservices, B-tree vs LSM, optimistic vs pessimistic locking, and 17 more.\n\n## Who this is for\n\nThis handbook is written for:\n\n- **SDE1 engineers preparing for SDE2 interviews.** Part 0 + Part 1 + Part 11, plus 8-10 targeted case studies from Part 8, plus the top-5 most-cited trade-off pages.\n- **SDE2s preparing for Senior\u002FStaff.** The full Parts 3-7 are the core \"you need to know this to design anything meaningful\" material. Parts 6 (Reliability) and 7 (Security) separate Senior candidates from Staff candidates.\n- **Senior\u002FStaff engineers refreshing or filling gaps.** Part 9 (AI\u002FML systems) and the Trade-offs Library are valuable even if you've been doing this for 15 years. Modern LLM serving and vector search weren't a thing when most of us learned backend.\n- **Self-taught engineers without a CS degree** who want the vocabulary and the reasoning that bootcamps and YouTube don't teach. Part 0 is explicit prerequisites: TCP\u002FIP, the OS contract, database internals, API design — fundamentals you need before \"distributed systems\" makes sense.\n- **Career switchers** moving from adjacent roles (frontend → backend, backend → infra, SWE → MLE) who need to build system-level intuition fast.\n- **Teachers and course creators** who want to build system-design curriculum without writing 800 pages from scratch. The CC BY-SA 4.0 license explicitly allows this.\n- **Anyone operating a production system at non-trivial scale** who wants a reference shelf that covers the full stack — from packet-level networking to multi-tenant SaaS isolation models.\n\n**This handbook is NOT for:**\n\n- **Absolute beginners with no programming experience.** Part 0 assumes you can read code and have built at least one CRUD app. If you're brand new, start with [The Odin Project](https:\u002F\u002Fwww.theodinproject.com\u002F) or [CS50](https:\u002F\u002Fcs50.harvard.edu\u002F), then come back.\n- **People who want leetcode\u002Falgorithms prep.** This handbook is HLD, not DSA. For algorithms, see [NeetCode](https:\u002F\u002Fneetcode.io\u002F) or [LeetCode](https:\u002F\u002Fleetcode.com\u002F).\n- **People who want low-level language-specific tutorials.** We don't teach Rust syntax or Go's goroutines — we teach concepts that apply across languages.\n\n## How the HLD Handbook compares\n\n|                                           | HLD Handbook | Typical OSS system-design repos | Paid courses (\\$150-\\$400+\u002Fyr) | O'Reilly-style books |\n| ----------------------------------------- | :----------: | :-----------------------------: | :---------------------: | :------------------: |\n| Free and open-source                      |      Yes     |                Yes              |            No           |           No         |\n| 100% inline content (no external redirects)|    Yes      |                No               |            Yes          |           Yes        |\n| 150+ chapters covering full stack         |      Yes     |                No               |      Only their sliver  |     Usually one topic|\n| 56 end-to-end case studies                |      Yes     |           5-15 typical          |         15-25 typical   |          Varies      |\n| Dedicated Trade-offs Library (22 pages)   |      Yes     |                No               |            No           |           No         |\n| Covers LLMs, RAG, agents, multimodal AI   |      Yes     |           Rarely modern         |         Sometimes       |         Rarely       |\n| Opinionated decisions, not just options   |      Yes     |                No               |        Sometimes        |         Varies       |\n| 3,100+ primary-source citations           |      Yes     |                No               |            No           |           Yes        |\n| Community-editable, PRs welcome           |      Yes     |                Yes              |            No           |           No         |\n| Updated continuously                      |      Yes     |       Often frozen              |        Varies           |      Every 3-5 years |\n| Decision flowcharts for architectural choices | Yes      |                No               |         Rarely          |         Rarely       |\n\n## Start reading\n\n> **For the best reading experience, visit [hld.handbook.academy](https:\u002F\u002Fhld.handbook.academy)** — free, no sign-up, full search, dark mode, and per-chapter diagram zoom. The links below open the source on GitHub, where diagrams also render natively.\n\nPick one:\n\n**I want a quick taste.** Read these three in order — they give you the vocabulary and the first real worked example:\n\n1. [Scalability: Growing a System Without Breaking It](content\u002Fpart-1-core-fundamentals\u002F00-scalability.md)\n2. [Back-of-the-Envelope Estimation](content\u002Fpart-1-core-fundamentals\u002F04-back-of-envelope-estimation.md)\n3. [Design a URL Shortener (TinyURL \u002F bit.ly)](content\u002Fpart-8-case-studies\u002F00-url-shortener.md)\n\n**I'm preparing for an SDE2 interview in the next 6 weeks.** Follow the [SDE1 → SDE2 study plan](#sde1--sde2--6-week-interview-prep).\n\n**I'm a Senior engineer refreshing my distributed-systems foundations.** Read [Part 3 — Distributed Systems Theory](#part-3--distributed-systems-theory-11-chapters) straight through.\n\n**I want to learn AI-systems design.** Read [Part 9 — AI & ML System Design](#part-9--ai--ml-system-design-15-chapters) + the 8 AI case studies in [Part 8](#part-8--case-studies-56-chapters) (chapters 30-37).\n\n**I want to read the whole book.** Start at [Part 0](#part-0--prerequisites-5-chapters) and move through the parts in order. At your target reading pace of one 25-minute chapter per day, you finish in about 6 months. Bookmark [hld.handbook.academy](https:\u002F\u002Fhld.handbook.academy) so your progress is one tap away.\n\n---\n\n## The full curriculum\n\n181 pages, every one linked below. Click any title to read the chapter on GitHub — Mermaid diagrams render natively.\n\n### Part 0 — Prerequisites (5 chapters)\n\n> **Audience:** engineers without a CS degree, or anyone who wants to confirm their foundations before moving on.\n> **Difficulty:** Beginner. **Total reading time:** ~3 hours.\n\nFoundational topics that the rest of the handbook assumes. If you can explain TCP's three-way handshake, the difference between a process and a thread, a B-tree's internal structure, and why idempotent PUT is better than non-idempotent POST for a retry-prone endpoint — you can skip this part.\n\n1. [Networking Fundamentals for System Design](content\u002Fpart-0-prerequisites\u002F00-networking-fundamentals.md)\n   — OSI and TCP\u002FIP layers, TCP vs UDP, HTTP\u002F1.1 vs HTTP\u002F2 vs HTTP\u002F3, DNS resolution, TLS handshake, what \"the network is unreliable\" really means in practice.\n2. [Operating System Essentials for System Design](content\u002Fpart-0-prerequisites\u002F01-os-essentials.md)\n   — Processes vs threads, context switching cost, virtual memory, page cache, filesystem I\u002FO, epoll\u002Fkqueue\u002FIOCP, why `O_DIRECT` matters for databases.\n3. [Data Structures for Distributed Systems](content\u002Fpart-0-prerequisites\u002F02-data-structures-for-systems.md)\n   — Hash tables, B-trees, LSM-trees, skip lists, Bloom filters, tries, HyperLogLog, Count-Min Sketch, and when each shows up in real systems.\n4. [Database Fundamentals for System Design](content\u002Fpart-0-prerequisites\u002F03-database-fundamentals.md)\n   — Transactions, isolation levels (read-committed to serializable), indexes, query planners, joins, and why your ORM hides things from you that you need to see.\n5. [API Design Basics: REST, GraphQL, gRPC, and the Hard Parts](content\u002Fpart-0-prerequisites\u002F04-api-design-basics.md)\n   — Resource modeling, idempotency, versioning, pagination, rate-limit headers, error envelopes, HATEOAS in theory vs practice.\n\n### Part 1 — Core Fundamentals (7 chapters)\n\n> **Audience:** everybody — read this part even if you know the topic, because the vocabulary here is used for the rest of the book.\n> **Difficulty:** Beginner-Intermediate. **Total reading time:** ~4 hours.\n\nThe vocabulary and the reasoning habits that every later chapter assumes. \"Scalability,\" \"consistency,\" \"trade-off,\" and \"back-of-envelope\" get defined here rigorously so they mean something specific when we use them later.\n\n1. [Scalability: Growing a System Without Breaking It](content\u002Fpart-1-core-fundamentals\u002F00-scalability.md)\n2. [Latency and Throughput: The Two Numbers That Matter](content\u002Fpart-1-core-fundamentals\u002F01-latency-and-throughput.md)\n3. [Availability and Reliability: Nines, SLOs, and Staying Up](content\u002Fpart-1-core-fundamentals\u002F02-availability-and-reliability.md)\n4. [Consistency Models: What Readers Actually See](content\u002Fpart-1-core-fundamentals\u002F03-consistency-models.md)\n5. [Back-of-the-Envelope Estimation](content\u002Fpart-1-core-fundamentals\u002F04-back-of-envelope-estimation.md)\n6. [How to Approach a System Design Question](content\u002Fpart-1-core-fundamentals\u002F05-how-to-approach-design-questions.md)\n7. [Trade-off Thinking](content\u002Fpart-1-core-fundamentals\u002F06-trade-off-thinking.md)\n\n### Part 2 — Building Blocks (16 chapters)\n\n> **Audience:** anyone who'll assemble backend systems. This part is the Lego brick inventory.\n> **Difficulty:** Intermediate. **Total reading time:** ~10 hours.\n\nDeep dives on the pieces you assemble to build real systems: load balancers, caches, queues, CDNs, databases (SQL and NoSQL), partitioning, replication, pub\u002Fsub, rate limiters, service discovery, blob storage, geospatial indexes, and edge compute.\n\n1. [Load Balancers: Spreading Traffic, Absorbing Failure](content\u002Fpart-2-building-blocks\u002F00-load-balancers.md)\n2. [Reverse Proxies and API Gateways: The Smart Edge](content\u002Fpart-2-building-blocks\u002F01-reverse-proxies-api-gateways.md)\n3. [Content Delivery Networks: Moving Bytes Closer to Users](content\u002Fpart-2-building-blocks\u002F02-cdns.md)\n4. [Caching: From Browser to Database](content\u002Fpart-2-building-blocks\u002F03-caching.md)\n5. [SQL Databases: The Boring Technology That Wins](content\u002Fpart-2-building-blocks\u002F04-sql-databases.md)\n6. [NoSQL Databases: Picking the Right Non-Relational Tool](content\u002Fpart-2-building-blocks\u002F05-nosql-databases.md)\n7. [Database Partitioning and Sharding: When One Node Is Not Enough](content\u002Fpart-2-building-blocks\u002F06-database-partitioning-sharding.md)\n8. [Database Replication: Keeping Copies in Sync](content\u002Fpart-2-building-blocks\u002F07-database-replication.md)\n9. [Message Queues and Streaming: Decoupling at Scale](content\u002Fpart-2-building-blocks\u002F08-message-queues-streaming.md)\n10. [Pub\u002FSub: Fan-Out and Event-Driven Systems](content\u002Fpart-2-building-blocks\u002F09-pub-sub.md)\n11. [Real-Time Communication: WebSockets, SSE, and Long Polling](content\u002Fpart-2-building-blocks\u002F10-real-time-communication.md)\n12. [Rate Limiting: Protecting Systems from Themselves](content\u002Fpart-2-building-blocks\u002F11-rate-limiting.md)\n13. [Service Discovery and Service Mesh: Finding and Talking to Services](content\u002Fpart-2-building-blocks\u002F12-service-discovery-mesh.md)\n14. [Blob and Object Storage: Storing the Big Stuff](content\u002Fpart-2-building-blocks\u002F13-blob-object-storage.md)\n15. [Geospatial Indexing: Geohash, Quadtree, R-tree, S2, and H3](content\u002Fpart-2-building-blocks\u002F14-geospatial-indexing.md)\n16. [Edge Computing (Cloudflare Workers, Lambda@Edge, Deno Deploy)](content\u002Fpart-2-building-blocks\u002F15-edge-computing.md)\n\n### Part 3 — Distributed Systems Theory (11 chapters)\n\n> **Audience:** SDE2+ preparing for Senior\u002FStaff. If you haven't internalized linearizability vs serializability, read this.\n> **Difficulty:** Intermediate-Advanced. **Total reading time:** ~9 hours.\n\nThe theory that makes distributed systems *distributed*: consensus (Raft\u002FPaxos), the full consistency spectrum, CAP\u002FPACELC in 2025 framing, logical clocks, CRDTs, distributed transactions (2PC\u002FSaga), exactly-once delivery, failure detection, consistent hashing, and Merkle-tree anti-entropy.\n\n1. [Consensus Protocols: How Distributed Systems Agree](content\u002Fpart-3-distributed-systems-theory\u002F00-consensus-protocols.md)\n2. [Consistency Deep Dive: Linearizability, Serializability, and the Spectrum Between](content\u002Fpart-3-distributed-systems-theory\u002F01-consistency-deep-dive.md)\n3. [Quorums and Replication: The Math of R + W > N](content\u002Fpart-3-distributed-systems-theory\u002F02-quorums-and-replication.md)\n4. [CAP and PACELC: The Tradeoff That Keeps Confusing People](content\u002Fpart-3-distributed-systems-theory\u002F03-cap-and-pacelc.md)\n5. [Clocks and Ordering: Lamport, Vector, and Hybrid Logical Clocks](content\u002Fpart-3-distributed-systems-theory\u002F04-clocks-and-ordering.md)\n6. [CRDTs: Conflict-Free Replicated Data Types](content\u002Fpart-3-distributed-systems-theory\u002F05-crdts.md)\n7. [Distributed Transactions: 2PC, Saga, and When to Avoid Both](content\u002Fpart-3-distributed-systems-theory\u002F06-distributed-transactions.md)\n8. [Idempotency and Exactly-Once: The Honest Truth About Delivery Guarantees](content\u002Fpart-3-distributed-systems-theory\u002F07-idempotency-exactly-once.md)\n9. [Failure Detection: Deciding a Node Is Dead](content\u002Fpart-3-distributed-systems-theory\u002F08-failure-detection.md)\n10. [Consistent Hashing: Keys to Nodes Without Global Reshuffles](content\u002Fpart-3-distributed-systems-theory\u002F09-consistent-hashing.md)\n11. [Merkle Trees and Anti-Entropy: Keeping Replicas in Sync Cheaply](content\u002Fpart-3-distributed-systems-theory\u002F10-merkle-trees-anti-entropy.md)\n\n### Part 4 — Data Systems (10 chapters)\n\n> **Audience:** anyone who owns a data pipeline or picks a database.\n> **Difficulty:** Intermediate-Advanced. **Total reading time:** ~8 hours.\n\nEvery flavor of data system you might pick, and when each actually fits. Storage engines (B-tree vs LSM), OLTP vs OLAP, warehouses\u002Flakes\u002Flakehouses, streaming vs batch, CDC, search, time-series, graph, vector, and key-value.\n\n1. [Storage Engines: B-Trees, LSM-Trees, and Why Your Database Feels the Way It Does](content\u002Fpart-4-data-systems\u002F00-storage-engines.md)\n2. [OLTP vs OLAP: Row Stores, Column Stores, and Matching Shape to Workload](content\u002Fpart-4-data-systems\u002F01-oltp-vs-olap.md)\n3. [Data Warehouses and Data Lakes: Structure, Schema, and the Lakehouse](content\u002Fpart-4-data-systems\u002F02-data-warehouses-lakes.md)\n4. [Stream vs Batch Processing: Lambda, Kappa, and the End of That Debate](content\u002Fpart-4-data-systems\u002F03-stream-vs-batch.md)\n5. [Change Data Capture: Streaming the Database's Inner Monologue](content\u002Fpart-4-data-systems\u002F04-change-data-capture.md)\n6. [Search Systems: Inverted Indexes, BM25, and Running Elasticsearch in Production](content\u002Fpart-4-data-systems\u002F05-search-systems.md)\n7. [Time-Series Databases: Metrics, Events, and Retention at Scale](content\u002Fpart-4-data-systems\u002F06-time-series-databases.md)\n8. [Graph Databases: Property Graphs, Cypher, and When Joins Are the Problem](content\u002Fpart-4-data-systems\u002F07-graph-databases.md)\n9. [Vector Databases: Embeddings, ANN Indexes, and the Retrieval Layer for AI](content\u002Fpart-4-data-systems\u002F08-vector-databases.md)\n10. [Key-Value Stores: Redis, Memcached, DynamoDB, and Picking the Right Hash Table](content\u002Fpart-4-data-systems\u002F09-key-value-stores.md)\n\n### Part 5 — Architecture Patterns (11 chapters)\n\n> **Audience:** engineers making architecture-level decisions or leading service migrations.\n> **Difficulty:** Intermediate. **Total reading time:** ~8 hours.\n\nThe architectural shapes you choose between when you design anything bigger than a single service: monolith vs microservices, event-driven, CQRS, event sourcing, serverless, BFF, strangler fig, hexagonal\u002Fclean, multi-region, multi-tenancy, CRDT-based apps.\n\n1. [Monolith vs Microservices: Team Topology, Conway's Law, and the Distributed System Tax](content\u002Fpart-5-architecture-patterns\u002F00-monolith-vs-microservices.md)\n2. [Event-Driven Architecture: Notifications, State Transfer, and Choreography](content\u002Fpart-5-architecture-patterns\u002F01-event-driven-architecture.md)\n3. [CQRS: Separating Reads from Writes Without Losing Your Mind](content\u002Fpart-5-architecture-patterns\u002F02-cqrs.md)\n4. [Event Sourcing: Events as the Source of Truth](content\u002Fpart-5-architecture-patterns\u002F03-event-sourcing.md)\n5. [Serverless: Functions, Cold Starts, and When FaaS Actually Saves Money](content\u002Fpart-5-architecture-patterns\u002F04-serverless.md)\n6. [Backend for Frontend: Per-Client API Aggregation Done Right](content\u002Fpart-5-architecture-patterns\u002F05-backend-for-frontend.md)\n7. [Strangler Fig: Incremental Migration Without a Big Bang](content\u002Fpart-5-architecture-patterns\u002F06-strangler-fig.md)\n8. [Hexagonal and Clean Architecture: Keeping Business Logic Independent](content\u002Fpart-5-architecture-patterns\u002F07-hexagonal-clean-architecture.md)\n9. [Multi-Region Architecture: Active-Passive, Active-Active, and CRDTs](content\u002Fpart-5-architecture-patterns\u002F08-multi-region-architecture.md)\n10. [Multi-Tenancy: Silo, Pool, and the SaaS Isolation Spectrum](content\u002Fpart-5-architecture-patterns\u002F09-multi-tenancy.md)\n11. [CRDT Applications (Yjs, Automerge, Local-First Software)](content\u002Fpart-5-architecture-patterns\u002F10-crdt-applications.md)\n\n### Part 6 — Reliability & Operations (11 chapters)\n\n> **Audience:** anyone on-call, anyone who signs SLAs, anyone paying a cloud bill.\n> **Difficulty:** Intermediate-Advanced. **Total reading time:** ~8 hours.\n\nThe engineering that separates \"it compiles\" from \"it runs reliably at 3 a.m. on a long weekend\": observability, SLOs, resilience patterns, auto-scaling, deployments, chaos engineering, incident response, health checks, FinOps, platform engineering.\n\n1. [Observability: Metrics, Logs, Traces, and the OpenTelemetry Standard](content\u002Fpart-6-reliability-and-operations\u002F00-observability.md)\n2. [SLI, SLO, SLA, and Error Budgets: Making Reliability Quantitative](content\u002Fpart-6-reliability-and-operations\u002F01-sli-slo-sla-error-budgets.md)\n3. [Resilience Patterns: Timeouts, Retries, Circuit Breakers, and Bulkheads](content\u002Fpart-6-reliability-and-operations\u002F02-resilience-patterns.md)\n4. [Graceful Degradation: When Partial Service Beats No Service](content\u002Fpart-6-reliability-and-operations\u002F03-graceful-degradation.md)\n5. [Auto-Scaling and Capacity Planning: From HPA to Predictive Scaling](content\u002Fpart-6-reliability-and-operations\u002F04-auto-scaling-capacity.md)\n6. [Deployment Strategies: Blue-Green, Canary, Rolling, and Feature Flags](content\u002Fpart-6-reliability-and-operations\u002F05-deployment-strategies.md)\n7. [Chaos Engineering: Breaking Things on Purpose](content\u002Fpart-6-reliability-and-operations\u002F06-chaos-engineering.md)\n8. [Incident Management: From Detection to Blameless Postmortem](content\u002Fpart-6-reliability-and-operations\u002F07-incident-management.md)\n9. [Health Checks and Readiness: Telling the Truth About Whether You're Up](content\u002Fpart-6-reliability-and-operations\u002F08-health-checks-readiness.md)\n10. [Cost Optimization and FinOps](content\u002Fpart-6-reliability-and-operations\u002F09-cost-optimization-finops.md)\n11. [Platform Engineering: IDPs, Golden Paths, and DX](content\u002Fpart-6-reliability-and-operations\u002F10-platform-engineering.md)\n\n### Part 7 — Security at Scale (10 chapters)\n\n> **Audience:** Senior\u002FStaff engineers, platform teams, security-adjacent builders.\n> **Difficulty:** Intermediate-Advanced. **Total reading time:** ~7 hours.\n\nSecurity architecture for real systems, not a CISSP crib sheet: AuthN vs AuthZ, OAuth2\u002FOIDC, JWT (and why you probably shouldn't), mTLS, secrets management, DDoS\u002FWAF, compliance (GDPR\u002FDPDP\u002FCCPA), software supply chain, privacy-preserving systems, post-quantum cryptography.\n\n1. [Authentication vs Authorization: Identity, Permissions, and Access Models](content\u002Fpart-7-security-at-scale\u002F00-authn-authz.md)\n2. [OAuth 2.0 and OpenID Connect: Delegated Authorization and Identity Done Right](content\u002Fpart-7-security-at-scale\u002F01-oauth2-oidc.md)\n3. [JWT Deep Dive: Signed Tokens, Claims, and the Revocation Problem](content\u002Fpart-7-security-at-scale\u002F02-jwt-deep-dive.md)\n4. [mTLS and Service-to-Service Authentication: SPIFFE, Service Mesh, and Zero Trust](content\u002Fpart-7-security-at-scale\u002F03-mtls-service-auth.md)\n5. [Secrets Management: Vault, KMS, and the End of Secrets in Config Files](content\u002Fpart-7-security-at-scale\u002F04-secrets-management.md)\n6. [DDoS Protection and WAFs: Mitigating Volumetric and Application Attacks](content\u002Fpart-7-security-at-scale\u002F05-ddos-waf.md)\n7. [Data Residency and Compliance Architecture (GDPR, DPDP, CCPA, Right-to-Erasure)](content\u002Fpart-7-security-at-scale\u002F06-data-residency-compliance.md)\n8. [Supply Chain Security: SBOM, SLSA, Sigstore, and Defending Against xz-utils](content\u002Fpart-7-security-at-scale\u002F07-supply-chain-security.md)\n9. [Privacy-Preserving Systems (Differential Privacy, Federated Learning)](content\u002Fpart-7-security-at-scale\u002F08-privacy-preserving-systems.md)\n10. [Post-Quantum Cryptography: Migrating to ML-KEM, ML-DSA, and a Crypto-Agile Future](content\u002Fpart-7-security-at-scale\u002F09-post-quantum-crypto.md)\n\n### Part 8 — Case Studies (56 chapters)\n\n> **Audience:** interview prep candidates, engineers building adjacent systems, anyone who learns best from worked examples.\n> **Difficulty:** Intermediate-Advanced. **Total reading time:** ~45 hours.\n\n56 end-to-end system designs, each following a consistent structure: requirements (functional + non-functional), back-of-envelope numbers, high-level architecture, data model, deep-dive components, scalability and reliability considerations, and real-world references. Grouped thematically below for easier navigation.\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Core primitives (chapters 00-03)\u003C\u002Fstrong> — the four designs you see in every interview\u003C\u002Fsummary>\n\n1. [Design a URL Shortener (TinyURL \u002F bit.ly)](content\u002Fpart-8-case-studies\u002F00-url-shortener.md)\n2. [Design a Pastebin (Paste Sharing Service)](content\u002Fpart-8-case-studies\u002F01-pastebin.md)\n3. [Design a Distributed Rate Limiter](content\u002Fpart-8-case-studies\u002F02-rate-limiter.md)\n4. [Design a Distributed Key-Value Store (Dynamo \u002F Cassandra \u002F Riak)](content\u002Fpart-8-case-studies\u002F03-key-value-store.md)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Messaging & social (chapters 04-09)\u003C\u002Fstrong> — notifications, chat, feeds, photos, crawlers, autocomplete\u003C\u002Fsummary>\n\n1. [Design a Notification System (Push, SMS, Email at Scale)](content\u002Fpart-8-case-studies\u002F04-notification-system.md)\n2. [Design a Chat System (WhatsApp \u002F Messenger \u002F Signal)](content\u002Fpart-8-case-studies\u002F05-chat-system.md)\n3. [Design a Social Media Feed (Twitter \u002F Instagram \u002F LinkedIn)](content\u002Fpart-8-case-studies\u002F06-social-media-feed.md)\n4. [Design a Photo Sharing Service (Instagram)](content\u002Fpart-8-case-studies\u002F07-photo-sharing.md)\n5. [Design a Web Crawler (Googlebot-style)](content\u002Fpart-8-case-studies\u002F08-web-crawler.md)\n6. [Design Search Autocomplete (Typeahead Suggestions)](content\u002Fpart-8-case-studies\u002F09-search-autocomplete.md)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Media & consumer products (chapters 10-17)\u003C\u002Fstrong> — video, ride-hailing, maps, file sync, editing, cache, recommenders\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n1. [Design a Video Streaming Service (YouTube \u002F Twitch \u002F TikTok)](content\u002Fpart-8-case-studies\u002F10-video-streaming.md)\n2. [Design Netflix (End-to-End)](content\u002Fpart-8-case-studies\u002F11-netflix.md)\n3. [Design a Ride-Hailing Service (Uber \u002F Lyft)](content\u002Fpart-8-case-studies\u002F12-ride-hailing.md)\n4. [Design Google Maps (Routing and Tile Rendering)](content\u002Fpart-8-case-studies\u002F13-google-maps.md)\n5. [Design a File Sync Service (Dropbox \u002F Google Drive)](content\u002Fpart-8-case-studies\u002F14-file-sync.md)\n6. [Design Collaborative Editing (Google Docs \u002F Figma \u002F Notion)](content\u002Fpart-8-case-studies\u002F15-collaborative-editing.md)\n7. [Design a Distributed Cache (Memcached \u002F Redis Cluster)](content\u002Fpart-8-case-studies\u002F16-distributed-cache.md)\n8. [Design a Recommendation System (Netflix \u002F YouTube \u002F TikTok)](content\u002Fpart-8-case-studies\u002F17-recommendation-system.md)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Commerce & financial (chapters 18-21)\u003C\u002Fstrong> — ticketing, payments, stock exchange, food delivery\u003C\u002Fsummary>\n\n1. [Design a Ticketing System (BookMyShow \u002F Ticketmaster)](content\u002Fpart-8-case-studies\u002F18-ticketing-system.md)\n2. [Design a Payment System (Stripe \u002F PayPal)](content\u002Fpart-8-case-studies\u002F19-payment-system.md)\n3. [Design a Stock Exchange (Matching Engine)](content\u002Fpart-8-case-studies\u002F20-stock-exchange.md)\n4. [Design a Food Delivery Service (DoorDash \u002F Swiggy)](content\u002Fpart-8-case-studies\u002F21-food-delivery.md)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Data & infrastructure (chapters 22-29)\u003C\u002Fstrong> — metrics, ad-click, logs, proximity, leaderboards, IDs, hotels, schedulers\u003C\u002Fsummary>\n\n1. [Design a Metrics Pipeline (Prometheus \u002F InfluxDB \u002F Thanos)](content\u002Fpart-8-case-studies\u002F22-metrics-pipeline.md)\n2. [Design Ad-Click Aggregation (Real-Time Stream Processing)](content\u002Fpart-8-case-studies\u002F23-ad-click-aggregation.md)\n3. [Design a Logging Platform (ELK \u002F Loki \u002F Splunk)](content\u002Fpart-8-case-studies\u002F24-logging-platform.md)\n4. [Design a Proximity Service (Nearby Friends \u002F Yelp)](content\u002Fpart-8-case-studies\u002F25-proximity-service.md)\n5. [Design a Real-Time Leaderboard](content\u002Fpart-8-case-studies\u002F26-leaderboard.md)\n6. [Design a Unique ID Generator (Snowflake, ULID, TSID, UUIDv7)](content\u002Fpart-8-case-studies\u002F27-unique-id-generator.md)\n7. [Design a Hotel Reservation System (Booking.com \u002F Airbnb)](content\u002Fpart-8-case-studies\u002F28-hotel-reservation.md)\n8. [Design a Distributed Job Scheduler (Airflow \u002F Temporal \u002F Distributed Cron)](content\u002Fpart-8-case-studies\u002F29-job-scheduler.md)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>AI systems (chapters 30-37)\u003C\u002Fstrong> — ChatGPT, RAG, coding agents, AI search, voice, moderation, semantic cache, model routing\u003C\u002Fsummary>\n\n1. [Design ChatGPT (Conversational AI at Scale)](content\u002Fpart-8-case-studies\u002F30-chatgpt-conversational-ai.md)\n2. [Design an Enterprise RAG System](content\u002Fpart-8-case-studies\u002F31-enterprise-rag.md)\n3. [Design a Coding Agent (Claude Code \u002F GitHub Copilot \u002F Cursor)](content\u002Fpart-8-case-studies\u002F32-coding-agent.md)\n4. [Design Perplexity (AI Search with Citations)](content\u002Fpart-8-case-studies\u002F33-perplexity-ai-search.md)\n5. [Design a Voice Agent (Alexa \u002F Siri-Class Realtime)](content\u002Fpart-8-case-studies\u002F34-voice-agent.md)\n6. [Design a Content Moderation System at Scale](content\u002Fpart-8-case-studies\u002F35-content-moderation-at-scale.md)\n7. [Design a Semantic Cache for LLM Applications](content\u002Fpart-8-case-studies\u002F36-semantic-cache.md)\n8. [Design a Model Router and Gateway (OpenRouter \u002F LiteLLM)](content\u002Fpart-8-case-studies\u002F37-model-router-gateway.md)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Infra services (chapters 38-39)\u003C\u002Fstrong> — feature flags, DNS\u003C\u002Fsummary>\n\n1. [Design a Feature Flag Service (LaunchDarkly \u002F Harness FME \u002F Unleash)](content\u002Fpart-8-case-studies\u002F38-feature-flag-service.md)\n2. [Design a DNS Service (Cloudflare 1.1.1.1 \u002F Google 8.8.8.8)](content\u002Fpart-8-case-studies\u002F39-dns-service.md)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Consumer products II (chapters 40-49)\u003C\u002Fstrong> — dating, auctions, SaaS, video conf, email, live comments, fraud, fitness, online judge, price tracking\u003C\u002Fsummary>\n\n1. [Design a Dating App (Tinder \u002F Hinge \u002F Bumble)](content\u002Fpart-8-case-studies\u002F40-dating-app.md)\n2. [Design an Online Auction (eBay \u002F Catawiki)](content\u002Fpart-8-case-studies\u002F41-online-auction.md)\n3. [Design a Multi-Tenant SaaS Platform](content\u002Fpart-8-case-studies\u002F42-multi-tenant-saas.md)\n4. [Design a Video Conferencing System (Zoom \u002F Google Meet)](content\u002Fpart-8-case-studies\u002F43-video-conferencing.md)\n5. [Design an Email Service at Gmail Scale (1.8B Users, 300B Messages\u002FDay)](content\u002Fpart-8-case-studies\u002F44-gmail-scale-email.md)\n6. [Design Live Comments at Scale (FB Live \u002F YouTube Live \u002F Twitch Chat)](content\u002Fpart-8-case-studies\u002F45-live-comments.md)\n7. [Design a Fraud Detection System (Stripe Radar \u002F PayPal \u002F Feedzai)](content\u002Fpart-8-case-studies\u002F46-fraud-detection.md)\n8. [Design a Fitness Tracking Service (Strava \u002F MapMyRun)](content\u002Fpart-8-case-studies\u002F47-strava-fitness.md)\n9. [Design an Online Judge (LeetCode \u002F Codeforces \u002F HackerEarth)](content\u002Fpart-8-case-studies\u002F48-online-judge.md)\n10. [Design a Price Tracking Service (CamelCamelCamel \u002F Honey \u002F Keepa)](content\u002Fpart-8-case-studies\u002F49-price-tracking.md)\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Developer & ops platforms (chapters 50-55)\u003C\u002Fstrong> — API gateway, CI\u002FCD, observability, search engine, brokerage, chat-at-scale\u003C\u002Fsummary>\n\n1. [Design an API Gateway at Scale (Kong \u002F AWS API Gateway \u002F Apigee \u002F Envoy)](content\u002Fpart-8-case-studies\u002F50-api-gateway.md)\n2. [Design a CI\u002FCD Platform (GitHub Actions \u002F GitLab CI \u002F CircleCI)](content\u002Fpart-8-case-studies\u002F51-cicd-platform.md)\n3. [Design an Observability Platform (Datadog \u002F New Relic \u002F Honeycomb)](content\u002Fpart-8-case-studies\u002F52-observability-platform.md)\n4. [Design a Search Engine (Google-Scale \u002F Brave Search)](content\u002Fpart-8-case-studies\u002F53-search-engine.md)\n5. [Design a Brokerage Platform (Robinhood \u002F E*TRADE \u002F Interactive Brokers)](content\u002Fpart-8-case-studies\u002F54-brokerage-trading.md)\n6. [Design Channel-Scale Chat (Discord \u002F Slack)](content\u002Fpart-8-case-studies\u002F55-channel-scale-chat.md)\n\n\u003C\u002Fdetails>\n\n### Part 9 — AI & ML System Design (15 chapters)\n\n> **Audience:** anyone building with or around LLMs, agents, or production ML.\n> **Difficulty:** Intermediate-Advanced. **Total reading time:** ~11 hours.\n\nModern AI-systems architecture, treated with the same rigor as Part 3: LLM serving, RAG, vector search, agent architectures, multi-agent orchestration, LLM evaluation, LLMOps, cost optimization, safety, ML fundamentals, feature stores, recommenders, real-time AI, multimodal, and the data infra underneath all of it.\n\n1. [LLM Serving Architecture (vLLM, TGI, TensorRT-LLM)](content\u002Fpart-9-ai-ml-system-design\u002F00-llm-serving-architecture.md)\n2. [RAG Pipelines (Retrieval-Augmented Generation)](content\u002Fpart-9-ai-ml-system-design\u002F01-rag-pipelines.md)\n3. [Vector Search at Scale (HNSW, IVF-PQ, DiskANN)](content\u002Fpart-9-ai-ml-system-design\u002F02-vector-search-at-scale.md)\n4. [AI Agent Architectures (ReAct, Reflection, Planning, Tool Use, Memory)](content\u002Fpart-9-ai-ml-system-design\u002F03-ai-agent-architectures.md)\n5. [Multi-Agent Orchestration (LangGraph, OpenAI Agents SDK, AutoGen, Swarm)](content\u002Fpart-9-ai-ml-system-design\u002F04-multi-agent-orchestration.md)\n6. [LLM Evaluation and Observability (Ragas, LangSmith, TruLens, LLM-as-Judge)](content\u002Fpart-9-ai-ml-system-design\u002F05-llm-evaluation-observability.md)\n7. [LLMOps and Prompt Engineering (Versioning, Guardrails, Red-Teaming)](content\u002Fpart-9-ai-ml-system-design\u002F06-llmops-prompt-engineering.md)\n8. [LLM Cost Optimisation (Semantic Cache, Model Routing, Cascading, Prompt Caching)](content\u002Fpart-9-ai-ml-system-design\u002F07-llm-cost-optimization.md)\n9. [LLM Safety and Guardrails (OWASP LLM Top 10, Prompt Injection, PII, Jailbreaks)](content\u002Fpart-9-ai-ml-system-design\u002F08-llm-safety-guardrails.md)\n10. [ML System Design Fundamentals](content\u002Fpart-9-ai-ml-system-design\u002F09-ml-system-design-fundamentals.md)\n11. [Feature Stores and Model Serving (Feast, Tecton, KServe, BentoML, MLflow)](content\u002Fpart-9-ai-ml-system-design\u002F10-feature-stores-model-serving.md)\n12. [Recommendation Systems Deep Dive (DLRM, Two-Tower, Embedding Retrieval, Cold Start)](content\u002Fpart-9-ai-ml-system-design\u002F11-recommendation-systems.md)\n13. [Realtime AI and Voice Agents (Streaming Inference, WebRTC, LiveKit, Deepgram)](content\u002Fpart-9-ai-ml-system-design\u002F12-realtime-ai-voice-agents.md)\n14. [Multimodal AI Systems (CLIP, Whisper, LayoutLM, Document AI)](content\u002Fpart-9-ai-ml-system-design\u002F13-multimodal-ai-systems.md)\n15. [Data Infrastructure for AI (Embedding Pipelines, Chunking, Unstructured ETL, MCP)](content\u002Fpart-9-ai-ml-system-design\u002F14-data-infrastructure-for-ai.md)\n\n### Part 10 — Emerging Patterns (1 chapter)\n\n> **Audience:** Staff+ engineers and architects thinking past 2026.\n> **Difficulty:** Intermediate-Advanced. **Total reading time:** ~45 minutes.\n\nForward-looking topics that are adjacent to everything else. Currently one chapter, with more planned (WebAssembly at the edge, unikernels, confidential computing, on-device AI).\n\n1. [Green Computing (Carbon-Aware Scheduling, PUE, Sustainable Systems)](content\u002Fpart-10-emerging-patterns\u002F00-green-computing.md)\n\n### Part 11 — Interview Framework (6 chapters)\n\n> **Audience:** anyone preparing for or giving system-design interviews.\n> **Difficulty:** Intermediate. **Total reading time:** ~4 hours.\n\nHow to run a 45-minute system-design interview, from both sides of the whiteboard. Compares RESHADED \u002F PEDALS \u002F ADEPT frameworks, teaches requirements scoping, diagramming, trade-off articulation, company-specific flavors, and RFC\u002Fdesign-doc authoring for Staff-level work.\n\n1. [Interview Frameworks Compared (RESHADED, PEDALS, ADEPT)](content\u002Fpart-11-interview-framework\u002F00-frameworks-compared.md)\n2. [Requirements Scoping: Functional, Non-Functional, and MoSCoW](content\u002Fpart-11-interview-framework\u002F01-requirements-scoping.md)\n3. [Diagramming Skills for System Design Interviews](content\u002Fpart-11-interview-framework\u002F02-diagramming-skills.md)\n4. [Trade-off Articulation: Saying 'It Depends' Well](content\u002Fpart-11-interview-framework\u002F03-trade-off-articulation.md)\n5. [Company-Specific Interview Flavors (Amazon, Google, Meta, Netflix)](content\u002Fpart-11-interview-framework\u002F04-company-specific-flavors.md)\n6. [Design Doc Authoring: RFCs, ADRs, and the Staff Engineer's Written Output](content\u002Fpart-11-interview-framework\u002F05-design-doc-authoring.md)\n\n### Trade-offs Library (22 pages)\n\n> **Audience:** everyone — these pages are referenced from every other part.\n> **Difficulty:** Intermediate. **Total reading time:** ~8 hours.\n\nThe 22 most-asked architectural-choice questions, each answered in a dedicated decision-comparison page: flowchart, comparison table, \"when to pick A\" vs \"when to pick B\" sections, real-world examples, and citations.\n\n1. [Strong vs Eventual Consistency](content\u002Ftrade-offs\u002F01-strong-vs-eventual-consistency.md)\n2. [ACID vs BASE](content\u002Ftrade-offs\u002F02-acid-vs-base.md)\n3. [SQL vs NoSQL](content\u002Ftrade-offs\u002F03-sql-vs-nosql.md)\n4. [Latency vs Throughput](content\u002Ftrade-offs\u002F04-latency-vs-throughput.md)\n5. [CAP and PACELC Applied](content\u002Ftrade-offs\u002F05-cap-and-pacelc.md)\n6. [Cache Strategies: Cache-Aside vs Write-Through vs Write-Behind](content\u002Ftrade-offs\u002F06-cache-strategies.md)\n7. [Batch vs Stream Processing](content\u002Ftrade-offs\u002F07-batch-vs-stream.md)\n8. [Load Balancer vs Reverse Proxy vs API Gateway](content\u002Ftrade-offs\u002F08-load-balancer-vs-proxy-vs-gateway.md)\n9. [REST vs gRPC vs GraphQL](content\u002Ftrade-offs\u002F09-rest-vs-grpc-vs-graphql.md)\n10. [Polling vs Long-Polling vs SSE vs WebSockets vs Webhooks](content\u002Ftrade-offs\u002F10-polling-vs-websockets.md)\n11. [Rate Limiting Algorithms: Token Bucket vs Sliding Window](content\u002Ftrade-offs\u002F11-rate-limiting-algorithms.md)\n12. [Optimistic vs Pessimistic Concurrency Control](content\u002Ftrade-offs\u002F12-optimistic-vs-pessimistic-locking.md)\n13. [Partitioning Schemes: Range, Hash, Consistent Hash, Directory](content\u002Ftrade-offs\u002F13-partitioning-schemes.md)\n14. [B-tree vs LSM-tree Storage](content\u002Ftrade-offs\u002F14-btree-vs-lsm.md)\n15. [Monolith vs Microservices](content\u002Ftrade-offs\u002F15-monolith-vs-microservices.md)\n16. [Replication Topologies: Leader-Follower, Multi-Leader, Leaderless](content\u002Ftrade-offs\u002F16-replication-topologies.md)\n17. [Distributed Transactions: 2PC vs Saga vs TCC](content\u002Ftrade-offs\u002F17-distributed-transactions.md)\n18. [Push vs Pull (Fan-out, Messaging, Feed)](content\u002Ftrade-offs\u002F18-push-vs-pull.md)\n19. [Lambda vs Kappa Architecture](content\u002Ftrade-offs\u002F19-lambda-vs-kappa.md)\n20. [Vertical vs Horizontal Scaling](content\u002Ftrade-offs\u002F20-vertical-vs-horizontal-scaling.md)\n21. [Normalization vs Denormalization](content\u002Ftrade-offs\u002F21-normalization-vs-denormalization.md)\n22. [Single-Region vs Multi-Region Deployment](content\u002Ftrade-offs\u002F22-single-vs-multi-region.md)\n\n---\n\n## Study plans\n\nYou don't need one — pick any chapter and start reading. These are here if you prefer a pre-built route, typically because you're preparing for a specific interview window or filling a specific gap.\n\n### SDE1 → SDE2 — 6-week interview prep\n\nGoal: pass a 45-60 minute system-design screen at a mid-to-senior level. Roughly 8-10 hours of reading per week.\n\n| Week  | Focus                       | Chapters                                                                              |\n| ----- | --------------------------- | ------------------------------------------------------------------------------------- |\n| **1** | Foundations                 | [Part 0](#part-0--prerequisites-5-chapters) + [Part 1](#part-1--core-fundamentals-7-chapters) (12 chapters, ~7 hrs) |\n| **2** | Building blocks I           | Part 2 chapters 0-7: load balancers, proxies, CDN, cache, SQL, NoSQL, partitioning, replication |\n| **3** | Building blocks II          | Part 2 chapters 8-15: queues, pub\u002Fsub, real-time, rate limiting, service mesh, blob storage, geo, edge |\n| **4** | Case studies (core)         | Pick 5 from Part 8 chapters 0-9: URL shortener, rate limiter, chat, feed, web crawler, autocomplete |\n| **5** | Case studies (your target)  | Pick 5 more from Part 8 relevant to your target company (see [Company-Specific Flavors](content\u002Fpart-11-interview-framework\u002F04-company-specific-flavors.md)) |\n| **6** | Interview mechanics         | [Part 11](#part-11--interview-framework-6-chapters) + top 5 most-cited pages in [Trade-offs Library](#trade-offs-library-22-pages) |\n\n### SDE2 → Senior — 3-month deep dive\n\nGoal: operate at Senior level, own cross-team architecture, pass loops at Senior+ bars.\n\n| Phase   | Focus                                           | Duration  |\n| ------- | ----------------------------------------------- | --------- |\n| **1**   | Foundations + building blocks                   | 3 weeks — [Parts 0-2](#part-0--prerequisites-5-chapters) (28 chapters) |\n| **2**   | Distributed theory + data + architecture        | 4 weeks — [Parts 3, 4, 5](#part-3--distributed-systems-theory-11-chapters) (32 chapters) |\n| **3**   | Case studies (all 56)                           | 4 weeks — [Part 8](#part-8--case-studies-56-chapters) |\n| **4**   | Reliability, security, AI, frontier, interview  | 2 weeks — [Parts 6, 7, 9, 10, 11](#part-6--reliability--operations-11-chapters) + [Trade-offs Library](#trade-offs-library-22-pages) |\n\n### Full curriculum — Staff+ preparation (6 months)\n\nRead everything in order. Use the end-of-chapter questions for active recall. Average one 25-minute chapter per day gets you through the whole thing in about 6 months, with buffer for harder chapters and re-reading.\n\n### AI\u002FML-only track (4 weeks)\n\nIf you already know the fundamentals and want to become fluent specifically in AI-systems design:\n\n| Week | Focus                | Chapters |\n| ---- | -------------------- | -------- |\n| 1    | LLM serving + RAG    | [Part 9](#part-9--ai--ml-system-design-15-chapters) chapters 0-3 |\n| 2    | Agents + evaluation  | [Part 9](#part-9--ai--ml-system-design-15-chapters) chapters 3-9 |\n| 3    | AI case studies      | [Part 8](#part-8--case-studies-56-chapters) chapters 30-37 (ChatGPT, RAG, coding agent, Perplexity, voice, moderation, semantic cache, model router) |\n| 4    | ML fundamentals      | [Part 9](#part-9--ai--ml-system-design-15-chapters) chapters 9-14 + [Recommendation System case study](content\u002Fpart-8-case-studies\u002F17-recommendation-system.md) |\n\n### Interview-triage track (one weekend)\n\nIf you have a loop on Monday:\n\n- Saturday morning: [How to Approach a System Design Question](content\u002Fpart-1-core-fundamentals\u002F05-how-to-approach-design-questions.md) + [Back-of-the-Envelope Estimation](content\u002Fpart-1-core-fundamentals\u002F04-back-of-envelope-estimation.md) + [Requirements Scoping](content\u002Fpart-11-interview-framework\u002F01-requirements-scoping.md) + [Diagramming Skills](content\u002Fpart-11-interview-framework\u002F02-diagramming-skills.md)\n- Saturday afternoon: 3 case studies in your domain\n- Sunday morning: 3 more case studies + [Trade-off Articulation](content\u002Fpart-11-interview-framework\u002F03-trade-off-articulation.md)\n- Sunday evening: [Company-Specific Interview Flavors](content\u002Fpart-11-interview-framework\u002F04-company-specific-flavors.md) for your target\n\n---\n\n## Project statistics\n\n| Metric                              | Count         |\n| ----------------------------------- | ------------- |\n| Parts                               | 12 + Trade-offs Library |\n| Teaching chapters                   | 159           |\n| Trade-off decision pages            | 22            |\n| **Total pages**                     | **181**       |\n| Total words                         | **~772,000**  |\n| Average words per chapter           | ~4,265        |\n| Mermaid diagrams                    | **719**       |\n| Citations to primary sources        | **3,100+**    |\n| Estimated total reading time        | ~110 hours    |\n| Equivalent printed book page count  | ~2,200 pages  |\n\nFor comparison: *Designing Data-Intensive Applications* is ~600 pages; Alex Xu's *System Design Interview* Vol. 1 is ~300 pages, Vol. 2 is ~340 pages; *Site Reliability Engineering* is ~550 pages. This handbook is longer than any three of them combined.\n\n## Quality standards\n\nEvery chapter in this repository passes 7 automated CI checks on every PR:\n\n1. **[markdownlint](https:\u002F\u002Fgithub.com\u002FDavidAnson\u002Fmarkdownlint)** — Markdown style and structure conformance.\n2. **[typos](https:\u002F\u002Fgithub.com\u002Fcrate-ci\u002Ftypos)** — source-code spell-check with a project-specific allowlist for technical terms.\n3. **Citation integrity** — every `[^1]`-style footnote has a corresponding `[^1]: source` definition; every citation is a real URL; no orphan citations.\n4. **Frontmatter validation** — every chapter declares `title`, `difficulty`, `prerequisites`, `date_created`, `date_updated`, `reading_time_minutes`, `tags` (from a canonical taxonomy), and `technologies` (from a curated allowlist).\n5. **Mermaid diagram validation** — all 719 diagrams must parse with `@mermaid-js\u002Fmermaid-cli` on CI so broken syntax doesn't ship.\n6. **[Vale](https:\u002F\u002Fvale.sh\u002F)** — prose-style linter for voice, passive voice, weasel words, and banned phrases. Custom rules in `.vale.ini`.\n7. **[lychee](https:\u002F\u002Fgithub.com\u002Flycheeverse\u002Flychee)** — external link checker run weekly; flags rotted URLs so citations stay valid.\n\nCI configuration lives in [`.github\u002Fworkflows\u002Fcontent-ci.yml`](.github\u002Fworkflows\u002Fcontent-ci.yml). Validator scripts are in [`scripts\u002F`](scripts\u002F).\n\nBeyond automation, every chapter is reviewed for:\n\n- Internal consistency (terminology matches Part 1 definitions)\n- Difficulty calibration (a \"Beginner\" chapter doesn't assume Staff-level context)\n- Diagram quality (Mermaid, not screenshots; captioned; accessibility-tagged)\n- Citation quality (primary sources preferred over summarizing blog posts)\n\n---\n\n## Contributing\n\n**Contributions of all sizes are welcome, from a typo fix to a full chapter.** Read [CONTRIBUTING.md](CONTRIBUTING.md) for the full workflow. A short summary:\n\n### Contribution paths\n\n| Time       | What you can do                                              | Issue required? |\n| ---------- | ------------------------------------------------------------ | :-------------: |\n| 5 min      | Fix a typo or dead link                                      | No              |\n| 15 min     | Add a missing citation or update an out-of-date number       | No              |\n| 30 min     | Add a real-world example or clarify a confusing paragraph    | No              |\n| 1 hour     | Create a Mermaid diagram for an existing chapter             | Optional        |\n| 2-4 hours  | Review a chapter for technical accuracy and leave feedback   | Yes             |\n| 4-8 hours  | Write a full chapter from an outline                         | **Yes, required** |\n| Translator | Translate one chapter or the full handbook into your language | Yes             |\n\n### How to contribute\n\n1. **Read the [STYLE_GUIDE.md](STYLE_GUIDE.md)** for voice, structure, diagram conventions, and citation format.\n2. **Open an issue first** for anything bigger than 30 minutes of work — this ensures you don't duplicate in-flight work.\n3. **Fork, branch, edit.** Use descriptive branch names like `fix\u002Fraft-quorum-math` or `add\u002Fmcp-protocol-chapter`.\n4. **Run validators locally (optional):**\n\n   ```bash\n   npm install\n   npm run check:all\n   ```\n\n   If you skip this, CI will run them on your PR anyway and tell you what to fix.\n5. **Submit a pull request** using the [PR template](.github\u002FPULL_REQUEST_TEMPLATE.md). Small fix? A one-sentence description is fine. Full chapter? Describe the pedagogical approach and list your primary sources.\n6. **Respond to review.** A maintainer will review within 7 days (usually faster). For content chapters, expect at least one round of technical review.\n\n### What makes a great contribution\n\n- **Correctness.** If you're citing a number, cite the primary source. If you're claiming a property (like \"Raft guarantees linearizability\"), cite the paper.\n- **Clarity.** Write for the difficulty tier declared in the chapter's frontmatter. Don't introduce Part-7 concepts in a Part-0 chapter.\n- **Opinion with evidence.** If you think the chapter should recommend something different, make the case with citations, not vibes.\n- **Pedagogical structure.** Intro → first principles → diagram → worked example → trade-offs → production gotchas → references. Deviate only when the topic genuinely demands it.\n\n### Reviewers wanted\n\nIf you have **operated one of the systems we cover in Part 8** (payment processing, real-time chat, feeds, video streaming, etc.) — your review is worth more than a month of solo research. Open an issue with your expertise area and which chapter you'd like to review.\n\n### Private\u002Fsensitive concerns\n\n- **Security issues in validator scripts or CI:** open a [private security advisory](https:\u002F\u002Fgithub.com\u002Fhandbook-academy\u002Fhld-handbook\u002Fsecurity\u002Fadvisories\u002Fnew).\n- **Code of conduct concerns:** email [hello@handbook.academy](mailto:hello@handbook.academy).\n- **Legal or licensing questions:** email [hello@handbook.academy](mailto:hello@handbook.academy).\n- **Direct contact:** [@invincible04](https:\u002F\u002Fgithub.com\u002Finvincible04).\n\n---\n\n## Project structure\n\n```text\nhld-handbook\u002F\n├── content\u002F                           # All 181 teaching pages (CC BY-SA 4.0)\n│   ├── part-0-prerequisites\u002F          # 5 chapters\n│   ├── part-1-core-fundamentals\u002F      # 7 chapters\n│   ├── part-2-building-blocks\u002F        # 16 chapters\n│   ├── part-3-distributed-systems-theory\u002F  # 11 chapters\n│   ├── part-4-data-systems\u002F           # 10 chapters\n│   ├── part-5-architecture-patterns\u002F  # 11 chapters\n│   ├── part-6-reliability-and-operations\u002F  # 11 chapters\n│   ├── part-7-security-at-scale\u002F      # 10 chapters\n│   ├── part-8-case-studies\u002F           # 56 chapters\n│   ├── part-9-ai-ml-system-design\u002F    # 15 chapters\n│   ├── part-10-emerging-patterns\u002F     # 1 chapter\n│   ├── part-11-interview-framework\u002F   # 6 chapters\n│   └── trade-offs\u002F                    # 22 decision pages\n│\n├── writing-guides\u002F                    # Author handbooks (CC BY-SA 4.0)\n│   ├── case-study-template.md         # Skeleton for Part 8 case studies\n│   └── trade-off-template.md          # Skeleton for trade-off decision pages\n│\n├── scripts\u002F                           # Content validators used by CI\n│   ├── check-citations.js             # Footnote integrity\n│   ├── check-frontmatter.js           # YAML frontmatter schema\n│   ├── check-mermaid.js               # Mermaid syntax validator\n│   ├── technologies.json              # Curated technology name taxonomy\n│   └── content-stats.js               # Word \u002F diagram \u002F citation counts\n│\n├── .github\u002F\n│   ├── workflows\u002F\n│   │   ├── content-ci.yml             # 8-job PR validation pipeline\n│   │   └── stale.yml                  # Issue\u002FPR hygiene\n│   ├── ISSUE_TEMPLATE\u002F                # Correction \u002F Request \u002F Writing templates\n│   ├── PULL_REQUEST_TEMPLATE.md\n│   ├── CODEOWNERS\n│   ├── FUNDING.yml\n│   ├── dependabot.yml\n│   └── release.yml\n│\n├── STYLE_GUIDE.md                     # Voice, structure, diagram conventions\n├── CONTRIBUTING.md                    # Full contribution workflow\n├── CODE_OF_CONDUCT.md                 # Contributor Covenant v2.1\n├── CONTRIBUTORS.md                    # All-contributors list\n├── SECURITY.md                        # Security disclosure process\n├── CITATION.cff                       # Citation metadata (Zenodo, academic tools)\n├── LICENSE                            # CC BY-SA 4.0 (everything in this repo)\n├── package.json                       # Dev dependencies for validators\n├── lychee.toml                        # Link-check configuration\n├── .vale.ini                          # Prose style rules\n├── .markdownlint.json                 # Markdown style rules\n└── .typos.toml                        # Spell-check allowlist\n```\n\n---\n\n## Development setup\n\nYou don't need anything installed to contribute content — edit Markdown, push, let CI validate. But if you want to run the checks locally before pushing:\n\n### Prerequisites\n\n- [Node.js 20+](https:\u002F\u002Fnodejs.org\u002F) (the `.nvmrc` file pins the exact version)\n- [npm](https:\u002F\u002Fwww.npmjs.com\u002F) (ships with Node)\n- Optional: [Vale](https:\u002F\u002Fvale.sh\u002F) for prose-style linting\n- Optional: [lychee](https:\u002F\u002Fgithub.com\u002Flycheeverse\u002Flychee) for link checking\n\n### Setup\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fhandbook-academy\u002Fhld-handbook.git\ncd hld-handbook\nnvm use                    # pins Node version from .nvmrc\nnpm install                # installs markdownlint, @mermaid-js\u002Fmermaid-cli, etc.\n```\n\n### Common commands\n\n```bash\n# Run all 7 content checks (same as CI)\nnpm run check:all\n\n# Individual checks\nnpm run lint               # markdownlint\nnpm run check:citations    # footnote integrity\nnpm run check:frontmatter  # YAML schema\nnpm run check:mermaid      # Mermaid diagram syntax\nnpm run check:typos        # spell-check\nnpm run check:links        # lychee link check (slow; runs weekly)\n\n# Optional\nnpm run prose              # Vale prose-style linter (requires Vale installed)\nnpm run stats              # word \u002F diagram \u002F citation counts\n```\n\n### Editor setup\n\nAny Markdown-aware editor works. Recommended extensions for VS Code:\n\n- `yzhang.markdown-all-in-one` — Markdown editing and preview\n- `bierner.markdown-mermaid` — Mermaid diagram preview\n- `davidanson.vscode-markdownlint` — inline markdownlint warnings\n- `chrischinchilla.vale-vscode` — Vale prose linter\n- `tekumara.typos-vscode` — typos spell-check\n\nThe repo includes `.editorconfig`, `.nvmrc`, `.markdownlint.json`, `.vale.ini`, and `.typos.toml` so your editor picks up project settings automatically.\n\n---\n\n## Governance\n\n**Benevolent maintainer model.** The project is currently solo-maintained by [@invincible04](https:\u002F\u002Fgithub.com\u002Finvincible04). Major architectural decisions (curriculum scope, licensing, CI strategy) are made by the maintainer after consultation with active contributors via Issues\u002FDiscussions.\n\n**Decision process:**\n\n- **Typo \u002F dead link \u002F factual correction:** any maintainer can merge.\n- **New chapter or major rewrite:** requires an issue with pedagogical rationale and at least one maintainer approval.\n- **Curriculum scope change (new part, renamed part, reordering):** requires a Discussion thread open for at least 7 days before merge.\n- **License change:** requires all-contributors consent; CC BY-SA 4.0 for everything in this repository is a stable commitment.\n\n**Becoming a maintainer.** Sustained, high-quality contributions over 3+ months can result in a maintainer invitation. This includes merge rights on your areas of expertise and a vote on curriculum-scope decisions.\n\nSee [CODEOWNERS](.github\u002FCODEOWNERS) for the current maintainer routing.\n\n---\n\n## FAQ\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Is this really free?\u003C\u002Fstrong>\u003C\u002Fsummary>\n\nYes, and specifically: everything in this repository — content, writing guides, style guide, validator scripts — is released under **CC BY-SA 4.0**. Anyone (including us) can share and adapt it, but adaptations must carry the same license. That means the chapter text cannot be paywalled by anybody — not by us, not by a future company, not by anyone who forks this repo. You can always read it free at [hld.handbook.academy](https:\u002F\u002Fhld.handbook.academy).\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Can I use this to build a paid course?\u003C\u002Fstrong>\u003C\u002Fsummary>\n\nYes. CC BY-SA 4.0 allows commercial use provided you:\n\n1. Attribute the original (\"Adapted from The HLD Handbook, CC BY-SA 4.0, https:\u002F\u002Fgithub.com\u002Fhandbook-academy\u002Fhld-handbook\")\n2. License your derivative under the same CC BY-SA 4.0 terms\n3. Indicate what you changed\n\nYou can teach courses, run boot camps, build YouTube channels, or sell books derived from this content — under share-alike. What you *can't* do is take the content, strip the attribution, and relicense it as proprietary.\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Can I translate it into my language?\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n**Please do.** Translations are explicitly welcomed. Open an issue saying which language you'd like to translate into and which chapters you plan to start with. Translations live in parallel directories (e.g. `content-hi\u002F` for Hindi, `content-zh\u002F` for Chinese) under the same repository, or as a linked sister repo — your call.\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>How is this different from `donnemartin\u002Fsystem-design-primer`?\u003C\u002Fstrong>\u003C\u002Fsummary>\n\n`system-design-primer` is a link-dump README that points at scattered blog posts and talks. It's great as a discovery tool. This handbook is inline content — every topic is taught fully within this repo. They're complementary: `system-design-primer` for breadth of resources, this handbook for depth of teaching.\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>How is this different from Alex Xu's books?\u003C\u002Fstrong>\u003C\u002Fsummary>\n\nAlex Xu's books are excellent and we cite them in nearly every case study. The differences:\n\n- **Scope:** ~640 pages (Vols 1+2) vs ~2,200 equivalent pages here.\n- **Freshness:** Xu Vol. 2 was published in 2022 and doesn't cover LLMs, RAG, agents, CRDTs, post-quantum crypto, MCP, etc. This handbook does.\n- **Format:** Xu is case-study-focused. This handbook has 56 case studies *plus* 103 foundational chapters that teach the building blocks.\n- **License:** Xu's books are copyrighted and paid. This handbook is CC BY-SA 4.0.\n- **Community:** Xu's books are one-author; this handbook accepts PRs.\n\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Is this the same as the website at hld.handbook.academy?\u003C\u002Fstrong>\u003C\u002Fsummary>\n\nThe **content** is identical. The website at [hld.handbook.academy](https:\u002F\u002Fhld.handbook.academy) adds a polished reading UI: full-text search across all 181 pages, dark mode, syntax-highlighted code blocks, per-chapter diagram zoom, social cards, OG images, and fast client-side navigation. Same content, better reading experience — and still free, with no sign-up.\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>Why don't you include an \"Awesome\" list of external resources?\u003C\u002Fstrong>\u003C\u002Fsummary>\n\nWe cite primary sources inline where they're relevant. A separate \"awesome\" list would duplicate that work and go stale faster. If a resource is worth reading, it's cited in a chapter's References section.\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>I want to submit a chapter on [topic X]. How?\u003C\u002Fstrong>\u003C\u002Fsummary>\n\nOpen an issue first with the [writing template](.github\u002FISSUE_TEMPLATE) explaining:\n\n1. Which part this chapter belongs in\n2. Why it belongs (not already covered? modern emerging topic? missing from Part X?)\n3. Proposed outline (1-2 paragraphs)\n4. Your primary sources\n\nA maintainer will respond with scope feedback within a week. Once approved, fork → draft → PR. For Part 8 chapters use [`writing-guides\u002Fcase-study-template.md`](writing-guides\u002Fcase-study-template.md); for trade-off pages use [`writing-guides\u002Ftrade-off-template.md`](writing-guides\u002Ftrade-off-template.md). All other chapters follow the skeleton documented in [STYLE_GUIDE.md](STYLE_GUIDE.md).\n\u003C\u002Fdetails>\n\n\u003Cdetails>\n\u003Csummary>\u003Cstrong>How do I cite this in an academic paper?\u003C\u002Fstrong>\u003C\u002Fsummary>\n\nUse the BibTeX entry in the [Citation](#citation) section below, or the `CITATION.cff` file (which academic tools like Zotero and Zenodo can parse directly).\n\u003C\u002Fdetails>\n\n---\n\n## License\n\nEverything in this repository — the 181 chapters, writing guides, style guide, validator scripts, templates, and configuration — is licensed under **[Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)](LICENSE)**.\n\n**What CC BY-SA 4.0 means in practice:**\n\nYou can:\n\n- Read it, share it, print it, screenshot it — forever.\n- Build courses, boot camps, YouTube channels, or books derived from it.\n- Translate it into any language.\n- Use it commercially.\n\nYou must:\n\n- **Attribute** the original (link back to this repo).\n- Release derivatives under the **same CC BY-SA 4.0 license** (share-alike).\n- Not relicense the content as proprietary or impose additional restrictions.\n\nNobody can paywall the chapter text itself under CC BY-SA 4.0. You can build anything on top of it, as long as the text itself stays open. Read the handbook free at **[hld.handbook.academy](https:\u002F\u002Fhld.handbook.academy)** — no sign-up, no paywall, ever.\n\n---\n\n## Citation\n\nIf you reference this project in academic work, blog posts, books, or courses:\n\n### BibTeX\n\n```bibtex\n@misc{hld-handbook,\n  author       = {Soni, Aayush},\n  title        = {The HLD Handbook: Open-Source High-Level Design Curriculum},\n  year         = {2026},\n  publisher    = {GitHub},\n  url          = {https:\u002F\u002Fgithub.com\u002Fhandbook-academy\u002Fhld-handbook},\n  note         = {159 chapters + 22 trade-off pages, 772K words, CC BY-SA 4.0}\n}\n```\n\n### Prose\n\n> *\"The HLD Handbook\" by Aayush Soni, https:\u002F\u002Fgithub.com\u002Fhandbook-academy\u002Fhld-handbook, CC BY-SA 4.0.*\n\n### Machine-readable\n\nThe [`CITATION.cff`](CITATION.cff) file in this repository is a [Citation File Format](https:\u002F\u002Fcitation-file-format.github.io\u002F) descriptor. GitHub's \"Cite this repository\" button, Zenodo, Zotero, and academic reference managers can parse it directly.\n\n---\n\n## Acknowledgments\n\nThis project stands on the shoulders of many, and we cite primary sources liberally throughout. A few that shaped the handbook most:\n\n**Canonical books:**\n\n- **Martin Kleppmann** — [Designing Data-Intensive Applications](https:\u002F\u002Fdataintensive.net\u002F) is the foundation underneath most of Parts 3-4.\n- **Alex Xu** — [System Design Interview](https:\u002F\u002Fwww.amazon.com\u002FSystem-Design-Interview-insiders-Second\u002Fdp\u002FB08CMF2CQF) Volumes 1 and 2 defined the interview case-study format we build on in Part 8.\n- **Chip Huyen** — [Designing Machine Learning Systems](https:\u002F\u002Fwww.oreilly.com\u002Flibrary\u002Fview\u002Fdesigning-machine-learning\u002F978109810795","HLD Handbook 是一本开源的高级设计与系统设计手册。该项目包含了181章内容、727千字和653张图表，涵盖了从基础知识到分布式系统理论、数据系统、架构模式等多个方面，并提供了丰富的案例研究和技术权衡分析。它采用JavaScript编写，并遵循CC BY-SA 4.0国际许可协议。HLD Handbook 非常适合软件工程师、架构师以及任何对系统设计感兴趣的人士使用，无论是为了提升个人技能还是准备技术面试，都是一个宝贵的学习资源。",2,"2026-05-29 04:05:13","CREATED_QUERY"]