Chapter 2 · Mechanics · lesson 4
Nodes and the mempool
8 min
Builds on Blocks and the chain
Chapter 2 · Mechanics · lesson 4
8 min
Builds on Blocks and the chain
"The network" is not an abstraction. It is thousands of ordinary computers, each running software that keeps a copy of the chain and refuses to accept anything that breaks the rules.
Each one is a node.
A node's job is unglamorous: receive blocks and transactions, verify every rule against its own copy, pass on what is valid, discard what is not.
There is no central list of participants. A node that decides to accept an invalid block does not corrupt the chain — it just disagrees with everyone else, and the rest of the network carries on without it.
This is what "decentralised" concretely means. Not that the data is spread around, but that the rules are enforced independently, thousands of times over, by parties who never agreed to trust each other.
Running a node is how you stop trusting anyone's answer about your balance and start checking it.
A transaction that has been broadcast but not yet included sits in each node's mempool. Two things about it surprise people:
It is not one shared pool. Every node keeps its own, and they differ. There is no canonical ordering and no fair queue.
It is public. Anyone can watch pending transactions before they are executed. Your intent is visible for seconds before it becomes real.
Predict
Validators pick from their mempool, and they are not obliged to be fair. Ordinary practice is to take the most profitable transactions, which usually means the highest tip.
Nothing forces a validator to include your transaction at all. A high fee is a strong incentive, not a guarantee.
Check