Chapter 1 · Foundations

CAS + CDC

Overview

Content-addressed storage (CAS) derives an object’s identity from its bytes. Equal bytes produce the same identity, so immutable objects can be verified and exact duplicates can share one stored copy. A whole-file CAS, however, loses that reuse when even a small edit changes the identity of the entire file.

Content-defined chunking (CDC) divides a file using boundaries derived from the content itself. A localized insertion, deletion, or replacement usually changes only nearby chunks; unchanged regions retain the same bytes and therefore the same CAS identities. Together, CAS and CDC represent a file as an ordered manifest of reusable, independently addressed chunks.

Combined path: file bytes → CDC boundaries → immutable chunks → CAS identities → ordered file manifest.