Sidetree

Sidetree solves a specific problem for anyone who wants to run a DID method without running a full blockchain of their own: how do you get a public, tamper-evident, globally-agreed record of "who controls this identifier and what key do they use now" without paying for your own consensus layer?

The answer is to batch. Instead of writing every DID create/update/deactivate operation as its own transaction on some chain, a Sidetree node collects a batch of operations, bundles them into a single file, and anchors just that file's hash with one small transaction on whatever underlying chain it's riding on (Bitcoin, in the best-known deployment). Anyone can independently fetch the batch file, verify it against the anchored hash, and replay the operations to reconstruct current DID state, all without the underlying chain knowing or caring that DIDs exist at all. That's the "Layer 2" framing: Sidetree borrows someone else's chain's security budget instead of building its own.

Where it comes from

Sidetree is a Decentralized Identity Foundation (DIF) specification, first proposed by Microsoft's Daniel Buchner and incubated inside DIF with major contributions from Transmute, SecureKey, Mattr, and Consensys over roughly four years of collaborative development, reaching DIF's approved v1 spec status in 2021. Microsoft's own implementation, ION, running on Bitcoin, is the best-known live deployment and is what most people picture when they hear the name. Solidus designed none of it and does not run it.

Solidus status

did:solidus deliberately does not use Sidetree, it took the opposite architectural bet on purpose, and it's worth naming the tradeoff both ways rather than only crediting one side. Instead of batching DID operations into a file anchored to someone else's chain, every did:solidus create, update, and revoke is itself a transaction that Solidus's own validator set finalizes directly, in real time.

Sidetree's advantage: a DID method can exist without ever running its own network, it rides on Bitcoin's (or another chain's) existing security, for as long as that chain exists. did:solidus's cost: it only resolves for as long as Solidus's own testnet is live; there is no borrowed security budget underneath it. And the tradeoff has cut against the Sidetree side too, not just in theory: Microsoft's own commercial product, Entra Verified ID, dropped did:ion as a supported trust system in favor of plain did:web, even though the open-source ION network itself continues to run, independent of that commercial decision. Neither path is free. Solidus is a DIF Associate Member, the same body that stewards Sidetree, but that membership is not an implementation of it and should never be described as one.

Proof: not applicable, Solidus doesn't implement Sidetree, so there's nothing of Solidus's to demonstrate. The honest side-by-side a stranger can run: resolve a did:ion DID through DIF's public ION resolver infrastructure, and resolve a did:solidus DID by calling solidus_didResolve directly against rpc.solidus.network, two different anchoring philosophies, answering the same kind of question in observably different ways.