Shifted by One Byte

A backup product stores nightly VM images that change only slightly between runs. The team is comparing fixed-size 4MB blocks against content-defined (variable-size) chunking for deduplication. Which statements are correct?

A

Shrinking the average chunk size raises the dedup ratio but grows the chunk index and per-chunk metadata cost

B

Deleting one backup can free its chunks immediately, since a chunk written by one night's run is owned by that run and referenced nowhere else

C

Content-defined boundaries survive an insertion near the start of the image, while fixed-size blocks re-cut every later block and lose their hashes

D

Unchanged regions hash to the same chunk identifiers on the next run, so only new chunks are uploaded and stored

E

Fixed-size blocks deduplicate better, because their boundaries are predictable and stay aligned with the disk sectors the VM image is written and read in