Fork data like you fork code. We shipped a new feature that lets you make a zero-copy fork of your bucket. Forks are a metadata only copy that create a new bucket based off a source bucket. It takes milliseconds to make a fork, and it doesn’t copy any data. We only save the data that has changed.
It’s fast, and it’s cheap. And it’s inspired by ZFS, as well as the append-only log based systems the Tigris team built at Uber.
You can read more about bucket forking here: Fork Buckets Like You Fork Code | Tigris Object Storage
And for the design philosophy, we wrote a blog post about immutability as a design principle: Append-Only Object Storage: The Foundation Behind Cheap Bucket Forking | Tigris Object Storage
Bucket forking is beautiful in context: for decades, storage has evolved to bring immutability layers up. Starting in the 90s with NetApp’s WAFL, then LVM, and ZFS, we’ve transformed data from mutable to immutable. Now Tigris is lifting immutability a layer up into object storage. We believe it will change how you think about data, just like how Git changed how you think about code.
Check it out! And as always, let us know what you think.