10TB featurs which stored on aws s3 bucket. in order to decrease costs , we are looking for a loseless compression method for those features (current features are images but it can changed between projects)
I heard about hdf5, but is it the best method nowadays? Can the the extraction be paralleled?