Relate or !Relate

on YouTube ↗53m

" summary

SourceForge replaces parts of its SQL backend with MongoDB to store project records and serve project summary, download and file pages; a four‑person team spent six weeks on the migration and about one week on the Mongo work. MongoDB yields surprising throughput (tested up to 18,000 writes/sec on a Mac Mini and tens of thousands of reads) and supports atomic partial updates and increment operators, but large project documents (one hit a 4 MB project record and saturated a gigabit link) required network tuning and careful data sizing.

AI-written summary · about

" chapters (6)

1. 0:00Start
2. 7:11ACID
3. 36:02taxonomy of NoSQL
4. 40:55partial updates
5. 41:40Conditional Updates
6. 43:25Figure out what

" related talks

• 48mGetting Non-Relational with MongoDBMichael Dirolf · RubyConf 2009
• 44mMongoDB to MySQL the How and the WhySarah Mei · RubyConf 2011
• 30mHelp! I Need To UnSQLize My ApplicationJoel Lord · PyCon Italia 2023
• 43mHelp! I Need To UnSQLize My ApplicationJoel Lord · Berlin Buzzwords 2022
• 18mFrom SQL to NOSQL - a gentle intro for DevsJoe Karlsson · Conf42 JavaScript 2020
• 51m5 boring performance improvements in MongoDB 3.4 that matterHenrik Ingo · HighLoad++ 2017
• 31mMongoDB: Advantages of an Open Source NoSQL DatabaseKevin Cearns · Spotlight: MEAN Stack 2014
• 29mSwitching to MongoDB: The challenges of translating database schemasKatie Bell · PyCon AU 2024