Home / Topics / Systems, Networks & Security
Distributed systems: videos Hacker News recommends
Distributed systems: consensus, Paxos, Raft, replication, CAP theorem, Kafka, event logs, Erlang/OTP, scaling services across machines, fault tolerance, large-scale architectures at Google, Netflix, Facebook. Below: every video in this topic that at least three HN users linked with real commentary, in at least three different months (66 talks, 5 lectures, 4 interviews), best-scored first, with the comments on each video page.
- It is HN's favorite pitch for Elixir and the BEAM, credited with convincing many people to adopt them.
- HN treats it as the canonical talk on simulation testing and a mind-blowing example of how far rigorous testing can go.
- The go-to talk showing how Erlang let a tiny team run a massive messaging system.
- Commenters repeatedly cite it as the talk that reframed how they think about caching, derived data and event sourcing.
- HN recommends it as a key Hickey talk on system design that changes how people think about distributed state and interfaces.
- The canonical recommendation for learning Go concurrency, praised as entertaining and clarifying when to use channels versus mutexes.
- Treated as the go-to approachable yet deep primer on CRDTs and collaborative editing, recommended repeatedly in threads on Automerge, local-first software and OT versus CRDT.
- HN treats it as the best introduction to Unison, with commenters finding hash-addressed immutable code mind-blowing.
- Diego Ongaro lecture recommended in 6 different months by 6 HN users, 2013–2024Recommended as a clear, lucid way to learn Paxos and Multi-Paxos, ideally alongside the companion Raft video.
- HN recommends it as the best video introduction to the actor model, straight from its inventor, and revisited it after Hewitt's death.
- HN regards it as one of the best technical talks, a deep and clear walkthrough of scalable lock-free hash tables built on compare-and-swap.
- Recommended as a fantastic introduction to declarative and Datalog-based distributed systems and failure modeling.
- Martin Kleppmann lecture recommended in 5 different months by 5 HN users, 2021–2025HN regularly recommends the series as a free, accessible, theory-oriented course on distributed computing that complements his book Designing Data-Intensive Applications.
- MIT 6.824: Distributed Systems lecture recommended in 6 different months by 5 HN users, 2020–2025Recommended as the standard free course for learning distributed systems from real papers.
- Recommended for its practical strategy for preserving state on ephemeral containers, and for the live demo.
- Groxio talk recommended in 5 different months by 4 HN users, 2023–2025HN recommends it as a rare real-world look at a large fault-tolerant Erlang/Elixir system with no downtime in healthcare.
- HN points people to it as Lamport's own case for thinking and writing a spec before coding, and as an entry point to TLA+.
- HN recommends it as a fascinating explanation of queues, flow control and overload that applies well beyond Clojure.
- HN treats Hamilton's annual talks as the best window into how AWS regions and hardware are actually built.
- Citus Data talk recommended in 5 different months by 4 HN users, 2016–2020HN commenters point to it as a first-hand overview of Heap's Postgres/Citus architecture and its distinctive indexing approach.
- Recommended as a detailed, candid case study of scaling a real system from zero users to millions.
- HN considers it one of the best game networking talks, with insights that carry over to distributed systems generally.
- HN recommends it as a transparent MMO postmortem and as the source of the story that the term 'shard' originated with Ultima Online.
- HN recommends it as an excellent talk on generalizing deterministic simulation testing, and Antithesis hiring posts link it to explain the work.
- HN people recommend it as a must-watch for anyone running complex systems, and some rewatch it regularly to remember its principles.
- Strongly recommended as a clear, accessible survey of the CRDT landscape.
- Recommended as an approachable introduction to CRDTs and the local-first idea behind apps like Actual.
- Reactive Summit talk recommended in 4 different months by 4 HN users, 2019–2020HN recommends it as an illuminating, approachable explanation of distributed consensus.
- Diego Ongaro lecture recommended in 4 different months by 4 HN users, 2013–2025HN commenters recommend it as a clearer introduction to Raft than the paper and as part of a well-designed understandability study.
- HN points to it as an accessible explanation of microkernel isolation and fault-tolerant design, and as Tanenbaum's answer to performance objections.
- HN recommends it as a classic look at systems engineering at Google, whose fundamentals have held up well despite its 2010 date.
- uncertainyesterday demo recommended in 4 different months by 4 HN users, 2013–2026A beloved classic that HN recommends as the quickest way to see why Erlang is special.
- HN engineers recommend it for its core insight that limiting concurrency naturally yields request-rate limits, and for its link to queueing theory.
- HN recommends it for getting the inventors of major concurrency models together, disagreements included.
- EMPEX Conference talk recommended in 3 different months by 3 HN users, 2017–2023Recommended as the best short overview of why the BEAM and Elixir stand apart from other runtimes.
- Martin Kleppmann talk recommended in 3 different months by 3 HN users, 2020–2024Recommended as an excellent first education on CRDTs and their limits.
- OnChain Scaling talk recommended in 3 different months by 2 HN users, 2017–2022Cited by big-block advocates as evidence that Bitcoin could handle far higher throughput.
- CH1LLW4VE demo recommended in 2 different months by 2 HN users, 2019–2021HN recommends it as a charming, concise demonstration of what makes Erlang special, from the people who built it.
- HN points to it as a simple diff-and-patch alternative to operational transforms and CRDTs for collaborative editing.
- Pointed to as the best source on Phoenix Presence internals and for combining Phoenix with Elm.
- Recommended as a concise, high-level intro to what TLA+ is and when it helps.
- Often brought up as an honest account of microservice sprawl, including the fact that nobody at Uber knew exactly how many services existed.
- netdevconf talk recommended in 2 different months by 2 HN users, 2023Recommended as the thorough case for replacing TCP in datacenters and for Ousterhout's Homa protocol.
- Perforce Puppet talk recommended in 2 different months by 2 HN users, 2017–2022HN recommends it for Hightower's clear mental models, such as comparing the scheduler to Tetris and explaining why resource limits matter.
- Pointed to as a rare look at how a tiny engineering team ran WhatsApp's massive scale on FreeBSD and Erlang.
- dawsdesign demo recommended in 2 different months by 2 HN users, 2017–2021HN treats it as a classic that shows what Erlang was built for: robust, concurrent telecom systems with live upgrades.
- Pointed to as evidence that Java can do fast, cheap message-passing concurrency, as an alternative to Erlang or Scala actors.
- Recommended as a compact preview of Kleppmann's data systems material, including from his book on the subject.
- Juan Benet demo recommended in 2 different months by 2 HN users, 2015–2017HN lists it among the best early resources for learning what IPFS is and does.
- Cited in decentralization and federation debates as the key talk on the practical challenges, though some found it defeatist.
- HN points to it as an authoritative overview of AWS infrastructure, including what AZs and regions really mean.
- Franklyn explainer recommended in 2 different months by 2 HN users, 2017–2018Commenters point to it as a short, high-level overview of how Lightning payment channels work.
- HN cites it as the authoritative description of Lambda internals and isolation, by the engineers who built it.
- Recommended as a conceptual rather than code-heavy talk on failure recovery from Erlang's creator.
- Linear talk recommended in 2 different months by 2 HN users, 2023–2025Recommended whenever local-first software, sync engines and CRDT alternatives come up, as a practical production account.
- Fraktio talk recommended in 2 different months by 2 HN users, 2021–2025Cited as the explanation of how Linear's fast, locally synced UI works.
- HN recommends it as a rich, concrete look at real-world system design and as evidence of YouTube's early Python stack.
- Pointed to as the canonical pitch for Erlang's fault-tolerance and self-healing model.
- Matrixdotorg talk recommended in 2 different months by 2 HN users, 2023–2024HN threads link it as the primary source for Matrix's performance claims and sliding sync improvements.
- Recommended as an entertaining cautionary talk on microservice failures and Conway's law.
- It is the go-to reference whenever HN discusses cell-based architecture and outage isolation, as an explanation straight from AWS engineers.
- Recommended as the protocol designer's own detailed explanation of SCP.
- HN points to it as an authoritative look at DynamoDB internals and as evidence it is a complex distributed system.
- Open Networking Summit talk recommended in 1 different month by 1 HN user, 2015–2017HN cites it as a rare public look at Google's internal network design and how it scales to many VMs.
- HN mostly cites it as evidence that YouTube was built on Python and as background on how it scaled.
- Mostly cited as a source for details of Cloudflare's architecture, such as LuaJIT-based request inspection and request volumes.
- Domain-Driven Design Europe talk recommended in 1 different month by 1 HN user, 2016–2026HN commenters cite it in event sourcing threads for Young's views on when to use it, including advice against cross-application event sourcing.
- Local-First Conf talk recommended in 1 different month by 1 HN user, 2025Cited in sync-engine and local-first threads for Kleppmann's argument for a generic, standardised sync protocol or service.
- HN points to it as insight into how AWS thinks about distributed systems and the difficulty of consensus.
- Cited as a useful taxonomy for pinning down what an event-driven design actually is.
- Open Networking Summit talk recommended in 1 different month by 1 HN user, 2013–2015Cited as a primary source for Google building its own switches and running OpenFlow in production.
- Cited in backup and data-loss threads as a rare inside look at Google's backup practices, including its use of tape.
- Cited as a technical look at how Tesla's VPPs and Autobidder work as distributed systems.
- Turing Awardee Clips interview recommended in 1 different month by 1 HN user, 2023–2024HN cites it as a primary source for Lamport's relativity inspiration behind logical clocks in distributed systems threads.
- HN cites it for insight into Google researchers' views on AI progress and publishing, though one commenter finds the interview weak.
- Confluent, an IBM Company talk recommended in 1 different month by 1 HN user, 2018–2020HN posts it in Kafka-as-database debates and as an introduction to event logs and CQRS.
- Mentioned as a real-world example that microservices can work at scale, though commenters are lukewarm and unsure it matches what they remember.
- Anyscale interview cited on HN 3 times, 2023–2023Cited as a source for the fact that OpenAI uses Ray for training, and for Brockman's views on developer speed.
- BSDTV talk cited on HN 5 times, 2014–2019Cited as the primary source for why Netflix uses FreeBSD and UFS rather than ZFS on its CDN boxes, and for its GPL reasoning.
- Google Workspace documentary cited on HN 5 times, 2017–2023Cited as a documentary look at how Google runs its data centers, though sometimes dismissed as PR.
- Cited as a real-world enterprise Java example, especially for JDK upgrades and ZGC results.
Questions
What is the most recommended distributed systems video on Hacker News?
The Soul of Erlang and Elixir • Sasa Juric • GOTO 2019 by GOTO Conferences: It is HN's favorite pitch for Elixir and the BEAM, credited with convincing many people to adopt them.
How is this distributed systems list ranked?
By HackerTube's score: the number of distinct months in which an HN comment recommended the video, plus half a point for every month with a strong recommendation. One viral thread counts less than steady recommendations over years.
Where do the videos come from?
From every Hacker News comment between 2007 and 2026 that linked a YouTube video, labeled by an AI model as a recommendation, a citation, a joke or a criticism. The list is published by HackerTube under ODC-By 1.0; this edition adds indexable pages and the comment context.