Leaves of Hash

Earlier in this post, we talked about Merkle trees as useful for commitment schemes and data integrity guarantees, but Miller et al. instead chooses to frame them as useful for delegation of data. If I have a Merkle tree with millions of elements in it, I can hand it over to a third party, retaining only the top hash, then make queries to this data expecting both a value and an inclusion proof. In Revocation Transparency, they introduce the notion of a sparse Merkle tree, a Merkle tree of infeasible size (in their example, depth 256, so a node per thousand atoms in the universe) where we exploit the fact that almost all leaves in this tree have the same value to compute proofs and digests in efficient time.

Source: blog.trailofbits.com