What 'layer two' means for Bitcoin

The Lightning specifications describe the protocol as a layer-two system for off-chain bitcoin transfer by mutual cooperation, relying on on-chain transactions for enforcement if necessary. A channel is established by a funding transaction that locks bitcoin in an output spendable only with both participants' signatures, and the participants then update the split between them by exchanging commitment transactions. Each new commitment revokes the previous one, so only the latest state can be enforced on chain.

Payments across more than one channel use a hashed timelock contract. The specification defines it as a conditional payment between two peers: the recipient can spend it by presenting a signature and a payment preimage, and the payer can cancel it after a given time. Because every hop along a route uses the same payment hash, a payment either completes at every hop or fails at every hop, and an intermediate node cannot take the funds. The specification notes that this is what allows channels to be chained into a network without trusting the intermediaries.

The trust model is the reason this counts as a layer two rather than a separate system. A channel participant is not trusting a third party with their bitcoin; they are relying on the base chain to enforce the latest signed state if the counterparty stops cooperating. The costs are real but different in kind: liquidity must be committed to channels in advance, a channel that closes unilaterally settles on chain and pays on-chain fees, and the timelock delays built into the protocol mean funds can be locked for a period during a disputed close.

Sidechains are not a single design. Liquid and Rootstock are both Bitcoin sidechains and they secure themselves in completely different ways, which is the clearest illustration that the category is defined by the peg rather than by the consensus.

Liquid uses a federation. Its documentation describes a fixed but upgradeable group of fifteen functionaries, drawn from exchanges, financial institutions and infrastructure companies, who collectively produce blocks, with each block requiring signatures from at least eleven of the fifteen. The same federation secures the peg: bitcoin is locked in a federation-controlled multisig wallet, and the documentation states that a peg-in requires 102 Bitcoin confirmations before the equivalent LBTC can be claimed, a requirement it explains as protection against deep reorganisations. Peg-out runs through a Peg-out Authorization Key system so that a compromised set of functionaries could not redirect funds to addresses of their choosing.

Rootstock takes the opposite approach to block production. Its developer documentation describes merged mining, in which the Rootstock chain is mined simultaneously with Bitcoin because both use the same double SHA-256 proof-of-work algorithm. A mining pool includes a commitment to the Rootstock block header in the Bitcoin coinbase transaction, and a solution that meets Bitcoin's difficulty also meets Rootstock's. The documentation is candid about the consequence: because Rootstock fees remain low relative to the Bitcoin block reward, the cost to attack Rootstock through double-spending is lower than Bitcoin's, and the project has built additional protections, including transparent double-spend trails and long miner reward maturity, to reduce that risk.

The two designs therefore fail differently. Liquid's security rests on a known set of institutions and their multisignature arrangements, which is fast and predictable but concentrates trust in a defined group. Rootstock's security rests on Bitcoin's hash rate, which is broad and expensive to acquire but which Rootstock only borrows, and which it cannot direct. Neither is a strictly better answer, and a reader choosing between them is choosing which assumption they find more comfortable.

The useful question is not whether a system is a layer two but where the bitcoin actually is and who can move it. If the bitcoin stays in an output on the Bitcoin chain that the holder can unilaterally spend after a timelock, the base layer is the backstop and the system inherits Bitcoin's settlement guarantees. If the bitcoin has been sent to an address controlled by a federation or a custodian, then the holder is relying on that group, and the security of the sidechain's consensus is a separate question from the security of the peg.

A second question is what happens when the operators stop cooperating. A payment channel has a defined unilateral exit: the holder broadcasts the latest commitment and waits out the timelock. A federated sidechain has a peg-out process that depends on the federation acting, and Liquid's documentation is explicit that general participants typically peg out indirectly through an exchange or a federation member. That is a materially different recovery path, and it is the one a holder should understand before committing funds.

The native and wrapped bitcoin page covers the third family of approaches, in which the asset is represented by a token issued against a custodian's holdings rather than by a peg into a chain with its own consensus. Read together, the two pages cover the full range of ways bitcoin is moved off the chain it was issued on, and the trust each one requires.