XRP Ledger Consensus Protocol

How the XRP Ledger reaches agreement on transactions without mining or traditional proof-of-stake.

The XRP Ledger does not use proof-of-work mining (like Bitcoin) or conventional stake-weighted voting (like most proof-of-stake chains). Instead, it uses the XRP Ledger Consensus Protocol (XRP LCP), a form of federated Byzantine agreement in which a network of independent validating servers repeatedly propose and vote on the next set of transactions until they converge on an agreed result.

Why not mining?

Proof-of-work requires validators to compete by burning enormous amounts of computation, which is energy-intensive and ties "who gets to decide the next block" to raw computing power. The XRPL's designers wanted transaction finality in seconds, not minutes, and wanted to avoid that energy cost entirely — so they built a voting-based alternative instead.

The basic process

Each round of consensus (producing one new "ledger version," roughly every 3–5 seconds) works approximately like this:

1. Collect. Each validator gathers the valid, pending transactions it has seen from the network (the candidate set).
2. Propose. Validators broadcast their proposed transaction set to their trusted peers.
3. Vote in rounds. Validators compare proposals from the validators they trust and update their own proposal to match the majority, repeating this over several rounds. In each round, the threshold of agreement required to keep a transaction in the set increases (starting around 50%, rising toward 80%).
4. Converge. Once at least 80% of a validator's trusted peers agree on the same transaction set, that set is considered to have reached consensus.
5. Apply and close. All nodes apply the agreed transactions in a canonical, deterministic order, compute the resulting ledger state, and "close" that ledger version — after which it is final and cannot be altered or reversed.

Which validators does a node trust?

Each server chooses its own list of trusted validators, called a Unique Node List (UNL). As long as the various UNLs used across the network overlap sufficiently and each validator behaves honestly, the network as a whole converges on a single, consistent transaction history. See Unique Node List and Validators for how trust is established and how this compares to a single global "member list."

What this buys the network

• Fast, deterministic finality — once a ledger closes, that's final; there's no analog to a blockchain reorg under normal operation.
• Low energy use — no computational race, so the marginal energy cost of consensus itself is small.
• No block rewards — because there's no mining race to incentivize, validators are not paid a block subsidy; see Transaction Fees and Burning for how the network's economics work instead.

What it requires in exchange

This design trades away permissionless, purely stake-weighted validator selection for reputation-based, opt-in trust. Anyone can run a validator, but for your validator's vote to matter to the rest of the network, other participants need to choose to trust it. This tradeoff — and the centralization criticism it has historically drawn — is discussed in more detail in The Centralization Debate.