Naming chunks by their hashes is not a new idea, but this technique along does not give you a practical backup tool. The deletion of unreferenced chunks become

Naming chunks by their hashes is not a new idea, but this technique along does not give you a practical backup tool. The deletion of unreferenced chunks becomes a hard problem, and the center piece of lock-free deduplication is the two-step fossil collection algorithm that solves this hard problem.