Execution
Execution is the full set of steps that carries out a financial action, such as a trade, a loan or a transfer, within or across networks, from request to a settled result. It may involve several transactions and unfold over time.
What it refers to
Execution refers to the process by which a financial action is actually carried out within and across networks.
It is not the same as a transaction.
A transaction is a discrete event recorded on a network. Execution is broader. It may involve several transactions, several networks, changes in liquidity, timing gaps and conditional steps.
Execution is about turning a request into a settled result. In practice, this could be a cross-network trade (often searched as cross chain swap), where an asset on one network is exchanged for an asset on another, or any action that needs coordinated steps to complete.
If a user wants to trade, lend, borrow or reposition assets, execution is the set of coordinated steps that carries that action through to the end.
It includes:
- Where the action happens
- How value moves
- What conditions must be met
- What state is reached at the end
Execution may unfold over time. It may involve more than one environment. It may complete in stages.
Why this concept exists
In simpler systems, execution and transaction were effectively the same thing.
Money lived in one place. A transaction was submitted. The result was immediate and final within that system.
That mental model breaks down in multi-network environments.
Today:
- Assets move across networks
- Liquidity differs by location
- Actions may depend on external conditions
- Finality may not be simultaneous everywhere
As a result, a single confirmed transaction no longer shows that the action the user asked for has been completed.
We need the concept of execution because the request, the transaction and the final result are no longer identical.
Execution names the full process that closes that gap.
What this changes for system design
If execution is broader than a transaction, systems cannot optimize only for transaction submission.
They must account for:
- Coordination across environments
- Time gaps between steps
- Conditions that may change during the process
- The difference between one step succeeding and the whole action completing
Execution becomes something that must be observed, guided and understood, not just triggered.
System design must therefore:
- Treat actions as processes rather than single events
- Make state transitions visible
- Acknowledge that completion may depend on more than one confirmation
- Provide a clear path back when an action cannot complete, the subject of recoverability
Execution shifts focus from "did a transaction go through?" to "did the action the user asked for actually complete?"
Related terms
- Cross-network execution: execution that spans networks
- Execution coordination: aligning the steps
- Recoverability: a clear path back when steps fail
- Asynchronous reality: why steps resolve at different times
- Cross-chain swap: a common example of execution