On: AWS switch from gzip to zstd – about 30% reduction...

[latchkey]

> There are APIs to fetch byte ranges from objects. Would that work without decompressing the entire object?

I wonder if you could store the objects in chunks and then individually compressed. Create an index of the byte ranges for those chunked objects so they can be looked up easily.

For example:

range 0-4000 -> objectA, objectB, objectC

Then just return those 3 objects, decompressed.