Banks and trading firms connected to the Canton network get to see and sign off on transactions that touch their own customers, even when those transactions are settled across shared infrastructure. The setup, described by the network's operators, splits transaction validation between the individual participant nodes and a common layer that coordinates how the pieces fit together.
Each participant node reviews the transactions involving its users before the shared infrastructure finalizes the result across the network. Product-access rules stay with the banks and venues themselves rather than being set at the network level.
Who checks what on Canton
The division of labor is the core of how Canton is pitched to regulated institutions. Participant nodes — the machines run by each bank or trading venue — are responsible for checking the transactions that involve their own users. That means a bank sees the activity it's on the hook for and can apply its own controls to it.
The shared infrastructure then coordinates the outcome of those transactions across the network. In other words, the network doesn't ask every participant to validate every trade; it relies on the parties with a direct stake in a transaction to do the checking, then synchronizes the result.
Separate rules for separate products
Banks and trading venues also set their own product-access rules inside the Canton network. A venue can decide who gets to trade a given product without waiting for a network-wide policy. That's a different model from a single public ledger where one set of rules applies to everyone by default.
The arrangement matters for institutions that have to answer to their own regulators and internal risk committees. Their access decisions stay under their own roof, while the plumbing that connects them to counterparties is shared.
What the split means for institutions
Canton is built for firms that want the efficiency of a common settlement layer but aren't willing to hand over customer-level oversight to a central operator. Keeping validation at the participant node keeps each firm's know-your-customer and transaction-monitoring obligations where they already sit.
The shared infrastructure handles the coordination step — making sure the separate validations agree on a single outcome. That's the part that has to work across institutions, and it's the part the network owns.
For banks evaluating whether to join, the practical question is how much of their existing compliance stack they can keep running at the node level versus what the shared layer requires them to change. The network's answer is that the node does the user-facing checks and the shared layer does the coordination.
The open question for participants
Product-access rules sitting with individual banks and venues means the network won't enforce a single eligibility standard across all participants. How those separate rules interact when a trade crosses two firms with different access policies is the detail institutions will want to pin down before they route significant volume through the network.
For now, the structure is set: participant nodes check their own users' transactions, shared infrastructure coordinates the outcome, and product access stays a local decision. Firms weighing participation will be testing how those three pieces hold together in practice.
