12 videos · 11 hr 32 minSystems and databases
These are the talks Hacker News points to when someone wants to know how real systems work underneath. It opens with a friendly story about scaling Dropbox. Then it moves through databases and storage, consensus and distributed data, running at scale, and performance. It ends with what happens when systems fail and how the people running them cope.
Play all on YouTube- 11:08:17
How We've Scaled Dropbox
Stanford · Picked by 5 · 2014–2023A good way in: a Dropbox engineer frankly walks through growing the service from nothing to millions of users, schema decisions included, and one commenter found it one of the most riveting things ever.
- 259:58
SQLite: The Database at the Edge of the Network with Dr. Richard Hipp
Method · Picked by 4 · 2017–2024The creator of the world's most widely used database explains where SQLite fits, how it is tested and why it gets long-term support, with an energy commenters enjoy as much as the content.
- 359:56
AWS re:Invent 2018: Amazon DynamoDB Deep Dive: Advanced Design Patterns for DynamoDB (DAT401)
Amazon Web Services · Picked by 12 · 2019–2025Rick Houlihan's dense session is HN's go-to for learning to design data around access patterns, and a clear overview of how NoSQL thinking differs from SQL.
- 447:43
"Turning the database inside out with Apache Samza" by Martin Kleppmann
Strange Loop Conference · Picked by 9 · 2015–2025Kleppmann reframes a database as a log of immutable facts with caches and indexes derived from it, the talk commenters say changed how they think about derived data.
- 5
1:06:34Paxos lecture (Raft user study)
Diego Ongaro · Picked by 6 · 2013–2024On to consensus: John Ousterhout's clear walkthrough of Paxos and Multi-Paxos, recorded for the Raft study and best watched alongside its Raft companion lecture.
- 6
51:36Conflict Resolution for Eventual Consistency • Martin Kleppmann • GOTO 2016
GOTO Conferences · Picked by 3 · 2017–2020When replicas disagree instead of agreeing on one leader, this approachable primer on operational transformation and CRDTs is what HN keeps linking in local-first threads.
- 7
40:20"Testing Distributed Systems w/ Deterministic Simulation" by Will Wilson
Strange Loop Conference · Picked by 12 · 2014–2024FoundationDB ran its whole database inside a simulator with injected faults to find rare bugs, and HN treats this as the defining talk on simulation testing.
- 8
49:15Erlang Factory 2014 - That's 'Billion' with a 'B': Scaling to the Next Level at WhatsApp
Erlang Solutions · Picked by 6 · 2014–2022Rick Reed shows how a handful of engineers ran WhatsApp's message load on FreeBSD and Erlang, with real server counts and the bottlenecks they fixed.
- 9
1:33:27AWS re:Invent 2016: Tuesday Night Live with James Hamilton
Amazon Web Services · Picked by 5 · 2016–2023Moving down to the physical layer, Hamilton tours how AWS builds its networking, power and custom silicon, which HN considers the best look inside a cloud provider.
- 10
59:34CppCon 2017: Matt Kulukundis “Designing a Fast, Efficient, Cache-friendly Hash Table, Step by Step”
CppCon · Picked by 9 · 2018–2026Performance at the scale of one data structure: the step-by-step story of SwissTable, the design that now sits behind Abseil, Rust's HashMap and Go's maps.
- 11
42:59"How NOT to Measure Latency" by Gil Tene
Strange Loop Conference · Picked by 9 · 2015–2026Before trusting any benchmark, learn why averages mislead and what coordinated omission is; some commenters call this talk life-changing.
- 12
52:35Debugging Under Fire: Keep your Head when Systems have Lost their Mind • Bryan Cantrill • GOTO 2017
GOTO Conferences · Picked by 8 · 2018–2025Cantrill recounts the day a mistyped command rebooted a whole Joyent datacenter, which HN remembers as the best postmortem it has seen, and a funny one.