Infrastructure

• From AI to sustainability, why our latest data centers use 400G networking

  // Nov 14, 2023

• How the data center site selection process works at Dropbox

  // Jun 13, 2023

• After four years of SMR storage, here's what we love—and what comes next

  // Mar 08, 2023

• Balancing quality and coverage with our data validation framework

  // Feb 07, 2023

• Increasing Magic Pocket write throughput by removing our SSD cache disks

  // Dec 08, 2022

• Future-proofing our metadata stack with Panda, a scalable key-value store

  // Nov 03, 2022

• Everything in its write place: Cloud storage abstraction with Object Store

  // Oct 11, 2022

• A look inside our sixth generation of server hardware

  // Jun 29, 2022

• That time we unplugged a data center to test our disaster readiness

  // Apr 25, 2022

• How we’re making Dropbox data centers 100% carbon neutral

  // Aug 03, 2021

• How we managed our global supply chain through a year of lockdown

  // May 20, 2021

• Boosting Dropbox upload speed and improving Windows’ TCP stack

  // May 03, 2021

• Atlas: Our journey from a Python monolith to a managed platform

  // Mar 04, 2021

• Lessons learned in incident management

  // Jan 05, 2021

• Alki, or how we learned to stop worrying and love cold metadata

  // Dec 17, 2020

• How we designed Dropbox ATF: an async task framework

  // Nov 11, 2020

• Keeping sync fast with automated performance regression detection

  // Aug 19, 2020

• Broccoli: Syncing faster by syncing less

  // Aug 04, 2020

• How we migrated Dropbox from Nginx to Envoy

  // Jul 30, 2020

• Engineering a disruption tolerant supply chain

  // Apr 28, 2020

• Testing sync at Dropbox

  // Apr 20, 2020

• Rewriting the heart of our sync engine

  // Mar 09, 2020

• Intelligent DNS based load balancing at Dropbox

  // Jan 08, 2020

• Evaluating BBRv2 on the Dropbox Edge Network

  // Dec 17, 2019

• Continuous integration and deployment with Bazel

  // Dec 11, 2019

• Monitoring server applications with Vortex

  // Nov 14, 2019

• Enhancing Bandaid load balancing at Dropbox by leveraging real-time backend server load information

  // Sep 18, 2019

• RunBMC: OCP hardware spec solves data center BMC pain points

  // Aug 21, 2019

• SMR: What we learned in our first year

  // Jul 30, 2019

• Your System is not a Sports Team

  // Jun 06, 2019

• Athena: Our automated build health management system

  // May 22, 2019

• How we optimized Magic Pocket for cold storage

  // May 06, 2019

• Embracing papercuts

  // Apr 17, 2019

• Don’t lead by example

  // Feb 21, 2019

• Finding Kafka’s throughput limit in Dropbox infrastructure

  // Jan 30, 2019

• The scalable fabric behind our growing data center network

  // Jan 23, 2019

• Automating Datacenter Operations at Dropbox

  // Jan 16, 2019

• Courier: Dropbox migration to gRPC

  // Jan 08, 2019

• Cape Technical Deep Dive

  // Dec 21, 2018

• Cross shard transactions at 10 million requests per second

  // Nov 09, 2018

• Dropbox traffic infrastructure: Edge network

  // Oct 10, 2018

• Building better compression together with DivANS

  // Jun 19, 2018

• Extending Magic Pocket Innovation with the first petabyte scale SMR drive deployment

  // Jun 12, 2018

• Meet Bandaid, the Dropbox service proxy

  // Mar 01, 2018

• Balancing open source and proprietary IP—they can co-exist

  // Dec 13, 2017

• Improving Document Preview Performance

  // Dec 01, 2017

• Deploying IPv6 in Dropbox Edge Network

  // Nov 09, 2017

• Handling system failures during payment communication

  // Sep 29, 2017

• Infrastructure update: evolution of the Dropbox backbone network

  // Sep 15, 2017

• Optimizing web servers for high throughput and low latency

  // Sep 06, 2017

• Evolution of Dropbox’s Edge Network

  // Jun 19, 2017

• Introducing Cape

  // May 17, 2017

• Deploying Brotli for static content

  // Apr 06, 2017

• Introducing Stormcrow

  // Mar 06, 2017

• Infrastructure Update: Pushing the edges of our global performance

  // Nov 16, 2016

• NetFlash: Tracking Dropbox network traffic in real-time with Elasticsearch

  // Oct 06, 2016

• Improving the performance of full-text search

  // Sep 07, 2016

• (Re)Introducing Edgestore

  // Aug 30, 2016

• Lepton image compression: saving 22% losslessly from images at 15MB/s

  // Jul 14, 2016

• Pocket watch: Verifying exabytes of data

  // Jul 06, 2016

• Lossless compression with Brotli in Rust for a bit of Pied Piper on the backend

  // Jun 29, 2016

• Going deeper with Project Infinite

  // May 24, 2016

• Enabling HTTP/2 for Dropbox web services: experiences and observations

  // May 11, 2016

• Inside the Magic Pocket

  // May 06, 2016

• Scaling to exabytes and beyond

  // Mar 14, 2016

• Project Schedule Estimation in Software Development

  // Oct 25, 2015

• Inside LAN Sync

  // Oct 13, 2015

• How to Write a Better Scribe

  // May 20, 2015

• Firefly: Instant, Full-Text Search Engine for Dropbox (Part 1)

  // Mar 22, 2015

• Dropbox at AWS re:Invent 2014

  // Dec 04, 2014

• Streaming File Synchronization

  // Jul 10, 2014

• Open Sourcing Our Go Libraries

  // Jul 01, 2014

• Introducing Pyston: an upcoming, JIT-based Python implementation

  // Apr 03, 2014

• Video Processing at Dropbox

  // Feb 18, 2014

• Improving Dropbox Performance: Retrieving Thumbnails

  // Jan 27, 2014

• Outage post-mortem

  // Jan 12, 2014

• Dropbox Status Update

  // Jan 10, 2014

• Caching in theory and practice

  // Oct 16, 2012

• Comtypes: How Dropbox learned to stop worrying and love the COM

  // Oct 04, 2012

• Dennis Ritchie

  // Oct 14, 2011