How does Ripple ledger work?

Ripple functions very differently than the other payment networks or cryptocurrencies. The Ripple network doesn’t have blocks, instead blocks are replaced by ledgers. Ripple provides a distributed ledger that logs all transactions.There is no mining involved in the Ripple network. In order to verify transactions the Ripple relies on voting consensus that is agreed upon by all nodes in the Ripple network. At its core, Ripple has a group of validators, each of which maintains its own Unique Node List (UNL).  UNL is a list of Ripple validators(node) that a given Ripple node believes are honest and will not double-cross them. Each Ripple node has their own list of trusted nodes.

When a Ripple transaction takes place it changes the state of ledger. Nodes receive unvalidated transactions, add them to their own current sets of candidates and pass them along to their connected validators. Validators only consider  transactions from their own UNL. The process of gathering unvalidated transactions and compiling them into a candidate set continues until the majority of the validators agree on the same set of transactions (candidate set) that will be added to the ledger. A node considers a candidate set as valid if 80% of the nodes on its UNL agrees it’s valid. Different groups of UNLs could reach 80% consensus individually with different sets of transactions. This would mean there will be no singular final version of ledger thus it will create a fork (chain split). To avoid this, the Ripple Labs established a default UNL of trusted validators that is run by Ripple themselves. Nodes are not required to connect to the default UNL, still the majority of the nodes choose to connect to the default UNL as it is difficult to create trusted UNLs and nodes who are syncing to the network for the first time are often isolated. The node with the largest connection that is densely interconnected with other nodes has the greatest influence on the Ripple network.