Reading up on data-storage
4 deep · digging since nov 25, 25
- SQLite Is a Library of Congress Recommended Storage Format
The US Library of Congress recommends SQLite as a preferred storage format for datasets, alongside XML, JSON, and CSV, prioritizing long-term readability and low patent risk.
- Files are the interface humans and agents interact with
Filesystems serve as a universal, durable database for both humans and AI agents, outliving the specific software used to create them.
- Inside The Internet Archive's Infrastructure
The article details the Internet Archive's custom PetaBox storage infrastructure, its engineering evolution since 2004, and the financial and operational challenges of preserving over 212 petabytes of web data.
- The unpowered SSDs in your drawer are slowly losing your data
Consumer SSDs left unpowered for over a year can lose data due to voltage leakage from NAND cells, making them risky for long-term cold storage.