Part 5 · The Network & Nodes

Bitcoin has no servers. It’s a gossip network of peers, each enforcing the rules independently. Consensus — the most-work chain we built in earlier parts — only means anything if thousands of strangers can independently learn about transactions, validate them, and agree on which block is the tip. This part is the plumbing that makes that possible: how data moves, who stores what, and where the real power sits.

The recurring thread of this whole book lands hardest here. How does a network of untrusting strangers agree on one ledger? The answer Part 5 makes concrete: every node verifies everything for itself, so nobody has to trust anybody. Miners propose blocks; nodes accept or reject them. The rules live in the software each person runs, not in any company or server.

Roadmap

We build the network from the participants up:

1. Node Types: Full, Pruned, SPV & Archival — what each kind of node validates and stores, and why a full node is your personal, sovereign verifier.
2. The P2P Protocol & Message Types — how peers find each other and speak: the version/verack handshake and the core message vocabulary.
3. Gossip: How Transactions & Blocks Propagate — the inv → getdata → tx/block flood, relay policy, and compact blocks.
4. The Mempool — each node’s private waiting room of unconfirmed transactions, and why there is no single global mempool.
5. Initial Block Download & Sync — how a brand-new node downloads and verifies the chain from genesis, trusting no one.
6. Network Attacks: Eclipse, Sybil & Partitioning — how the gossip layer can be attacked, and how nodes defend themselves.

The recurring question, applied here

How does the P2P network help untrusting strangers agree on one ledger? Every full node independently verifies every rule — supply, signatures, no double-spends, script validity — so no one has to trust a peer, a miner, or a server. They trust the rules, and check them themselves.