Bitcoin Network Rabbit Hole

Your transaction doesn’t go to a miner directly — it broadcasts to the nearest nodes on the Bitcoin network, which pass it to their neighbors, which pass it to theirs, spreading across the globe within seconds until every node and miner on earth has seen it.

Your phone talks to a node — a computer running Bitcoin software. That node checks the transaction is valid: your signature is genuine, you actually have the funds, the format is correct. Then it passes the transaction to every other node it’s connected to. Those nodes pass it to theirs. Within seconds it has traveled across the entire global network through a process called a gossip protocol — the same way rumors spread, except every step is mathematically verified.

Nobody “receives” your transaction the way an email has a recipient. It’s more like a public announcement that gets heard everywhere simultaneously. Every miner on earth now knows you want to send Bitcoin. They just haven’t included it in a block yet.

That gap between “heard everywhere” and “confirmed” is where things get interesting.

---

Every node on the Bitcoin network maintains its own copy of the mempool — a pool of valid but unconfirmed transactions — meaning your transaction lives in thousands of places simultaneously until a miner picks it up and locks it into a block.

The mempool isn’t a single place. It’s thousands of places at once. Every node runs its own local version. When your transaction arrives, each node independently validates it and adds it to their list. They all have slightly different mempools — nodes have memory limits and some drop low-fee transactions when congestion is high — but the overlap is enormous.

Miners watch the mempool constantly, scanning for the highest-paying transactions to include in their next block. Your transaction is sitting in plain sight, visible to every miner on earth, waiting to be chosen.

One thing worth knowing: the mempool is completely public. Anyone can look at it. Your transaction amount, the addresses involved, the fee you’re paying — all visible before a single confirmation. Bitcoin is pseudonymous, not anonymous. The mempool is where that distinction matters most.

How long it waits there comes down to one thing — and it’s entirely in your control before you send.

---

Transaction confirmation speed is determined almost entirely by the fee rate you paid — during busy periods, transactions with low fees can wait hours or even days in the mempool while higher-paying transactions jump the queue ahead of them.

Miners sort the mempool by feerate — fee per unit of transaction size, measured in satoshis per virtual byte. High feerate gets picked first. Low feerate waits.

During quiet periods, even very low fees confirm quickly because block space is plentiful. During busy periods — major market moves, waves of Ordinal inscriptions, anything that floods the mempool — block space becomes scarce and fees spike. If you sent with a low fee just before the network got busy, you’re stuck in the queue while everyone behind you who paid more jumps ahead.

The blocks themselves aren’t the variable — they still come roughly every 10 minutes. It’s the queue. If 50,000 transactions are waiting and each block holds a few thousand, the ones paying most get in first. The rest keep waiting.

Transactions don’t automatically expire in Bitcoin. They can sit in mempools for days or weeks. If a node restarts or clears its mempool during that time, your transaction might drop out and need to be rebroadcast by your wallet — though most wallets handle this automatically.

Which raises the practical question: how are you supposed to know what fee to pay before you send?

---

Most wallets estimate fees automatically by reading current mempool congestion, but the right fee depends on how urgently you need confirmation — and overpaying is common because wallets tend to be conservative.

Good wallets show you options: slow (cheap, could take hours), standard (confirmed within a few blocks), fast (next block or two, costs more). These estimates come from reading the current mempool — seeing what feerate is needed to get into the next few blocks given current congestion.

The complication: wallets are estimating a future state. If the network gets busier after you send, your “standard” fee might not be enough. If it quiets down, you overpaid. Fee markets are genuinely unpredictable in the short term.

The practical approach: use your wallet’s recommended fee for most transactions. For anything time-sensitive, bump it up. For anything that can wait, go low and be patient. Sites like mempool.space show you live mempool depth and fee estimates if you want to check conditions before sending.

There’s no universally “right” fee — only right for your urgency and the network’s current state.

Fees determine priority. But there’s a more fundamental question sitting underneath all of this — when you broadcast a transaction, who exactly sees it, and what can they see about you?

---

Miners see your transaction through the public mempool — they can see the amount, the addresses, and the fee, but they don’t see your name, and no identity information is attached to the transaction itself.

The transaction contains: the addresses the Bitcoin is coming from, the addresses it’s going to, the amount, and a cryptographic signature proving you authorized it. No name. No email. No identity document. The transaction is a mathematical object, not a personal communication.

That said — your IP address is visible to the nodes your wallet connects to when broadcasting. If you’re using a regular wallet without privacy tools, the first node you talk to knows your IP sent that transaction. Privacy-conscious users route Bitcoin through Tor or use their own node to remove this exposure.

The addresses themselves are pseudonymous — not tied to your identity unless you’ve revealed the connection elsewhere. Buying from an exchange that collected your ID, receiving Bitcoin at an address you posted publicly, or spending patterns that reveal behavior — all of these can link addresses to identities. Chain analysis companies exist specifically to trace these connections.

Bitcoin is not private by default. It’s pseudonymous by default — identity-free transactions on a fully transparent ledger. Understanding that distinction matters for how you use it.

The nodes receiving and validating your transaction before miners ever see it are doing a job that’s easy to underestimate. Which brings up a question most people never think to ask.

---

A node is any computer running Bitcoin software and maintaining a full copy of the blockchain — nodes validate and relay transactions and blocks, while miners are a specialized subset of nodes that also do the proof-of-work computation to create new blocks.

Every miner runs a node. But most nodes are not miners.

A full node downloads the entire Bitcoin blockchain — currently around 600GB — validates every transaction and block against Bitcoin’s rules independently, and maintains its own copy of the ledger. It doesn’t trust anyone. It checks everything itself. When a miner finds a new block, nodes verify it independently before accepting it. If a miner tried to cheat — creating Bitcoin from nothing, or spending someone else’s coins — every honest node on the network would reject the block immediately, regardless of how much mining power was behind it.

This is why nodes are Bitcoin’s real enforcement layer. Miners produce blocks. Nodes decide which blocks are valid. A miner with 90% of the world’s hash rate cannot change Bitcoin’s rules — because nodes running the original rules would simply reject the invalid blocks.

Running a node requires a regular computer, a broadband connection, and around 600GB of storage. The barrier is intentionally low. Mining requires industrial hardware and cheap electricity. The gap between those two things is important.

Which creates an interesting question about everyone who’s neither a miner nor a node operator.

---

Most Bitcoin users are light clients — they rely on wallets and apps that connect to other people’s nodes rather than verifying the blockchain independently, which is convenient but means trusting someone else’s copy of the ledger.

When you use Cash App, Coinbase, or most mobile wallets, you’re not running a full node. You’re connecting to servers that run nodes on your behalf — trusting that their version of the blockchain is honest and complete.

This is a tradeoff. Full nodes are self-sovereign — they verify everything themselves and trust nobody. Light clients are convenient — they check the parts that matter to you without downloading 600GB, but they do rely on the nodes they connect to being honest.

For most everyday users, this tradeoff is perfectly reasonable. For large amounts, or for people who want genuine financial sovereignty, running your own node removes that trust requirement entirely. Satoshi designed the system so that anyone who wanted to verify independently could — cheaply and easily. The existence of many independent verifiers is precisely what keeps the whole system honest.

Which raises something that seems counterintuitive at first.

---

Node operators earn no Bitcoin — they run nodes because doing so lets them verify their own transactions without trusting anyone else, and because a healthy network of independent nodes is what prevents any single entity from changing Bitcoin’s rules.

The incentive is sovereignty, not profit.

If you run a full node, you don’t need to trust an exchange, a wallet provider, or a miner to tell you what’s valid. You check it yourself against your own complete copy of history. For someone holding significant Bitcoin, that independence is worth more than any reward.

Beyond personal sovereignty, nodes are Bitcoin’s constitutional layer. They enforce the rules. More independent nodes means more distributed enforcement — harder for any government, company, or miner coalition to push through a rule change that the broader community doesn’t want. The 2017 block size war was ultimately decided by nodes: when a powerful coalition of miners tried to force a block size increase, nodes running the original rules simply rejected the new blocks. The miners backed down.

Running a node is a vote that never stops counting.

With all that said — how many of these votes actually exist, and are they distributed enough to matter?

---

Bitcoin has approximately 15,000–20,000 publicly reachable full nodes worldwide, with many more running privately — distributed across dozens of countries, making any coordinated shutdown by a single government effectively impossible.

The reachable count — nodes that accept incoming connections — sits around 15,000–20,000 at any time, visible on network explorers. Many more run behind home routers and firewalls without accepting incoming connections. Total full node estimates range from 50,000 upward.

Geographically they’re spread across the United States, Germany, France, the Netherlands, and dozens of other countries. No single country hosts more than roughly 20% of the visible network.

A government can ban Bitcoin and shut down nodes within its borders. China effectively did this during the 2021 mining ban — and Bitcoin kept running without interruption. The remaining nodes absorbed the load and the network didn’t miss a beat.

To meaningfully threaten the node network globally, you’d need coordinated action across dozens of jurisdictions simultaneously — including countries that are actively hostile to each other geopolitically. The distributed, multinational nature of Bitcoin’s infrastructure is a deliberate security feature.

The distribution handles political pressure. But there’s a more basic question about how a network with no central server stays coherent at all.

---

Bitcoin uses a peer-to-peer gossip protocol where every node connects to multiple peers simultaneously and shares new information with all of them — so data propagates redundantly across the network regardless of which individual nodes are online at any moment.

Each node maintains connections to 8–125 other nodes at the same time. When a new transaction or block arrives, the node passes it to every connected peer. Those peers pass it to theirs. The information spreads in seconds through overlapping, redundant paths.

There’s no single route a transaction must take. If 40% of nodes went offline right now, the remaining nodes would still propagate information through surviving connections. The network degrades gracefully — slower, not broken.

New nodes joining the network find peers through hardcoded seed nodes and a DNS-based discovery system. A node can go offline for months, come back, and catch up by downloading the blocks it missed from its peers. Nothing is lost. Everything is reconstructible.

The design deliberately has no critical path and no single point of failure. Every redundancy is intentional. Satoshi understood from the start that networks fail — so he built something that routes around failure rather than depending on reliability.

Redundancy handles nodes going offline. But what about someone deliberately trying to overwhelm the network?

---

Bitcoin’s fee market is its natural spam filter — flooding the mempool with transactions costs real money in fees, and since block space is limited, spam transactions simply sit in the queue or get dropped, without disrupting legitimate confirmed transactions.

This has been tried. In 2015 and again during the 2017 congestion period, attackers flooded the mempool with thousands of low-value transactions, bloating it and driving up fees for everyone. It worked — temporarily. Fees spiked. Confirmation times stretched. It was annoying.

But the attack is self-limiting in two ways. First, every transaction requires a fee — sustained spam requires sustained spending. The attacker has to keep paying. Second, when a node’s mempool gets too full, it automatically drops the lowest-feerate transactions first. Spam with tiny fees gets pruned.

The blockchain itself was unaffected throughout. Confirmed transactions — the ones paying adequate fees — continued confirming normally. The spam clogged the waiting room, not the road.

Making block space scarce and priced isn’t just an economic mechanism. It’s a defense mechanism.

Spam attacks aside — there’s a more personal version of this stuck-transaction problem that many users have actually faced.

---

You can’t cancel a confirmed Bitcoin transaction, but if it’s still unconfirmed in the mempool, you can replace it with a higher-fee version using a feature called Replace-by-Fee (RBF) — essentially outbidding your original transaction for miner attention.

Once a transaction is confirmed in a block, it’s permanent. No cancellation, no reversal, no chargeback. This is fundamental to what Bitcoin is — a final settlement layer. The irreversibility is the feature.

But while it’s still unconfirmed — floating in the mempool — you have options if your wallet supports RBF. You broadcast a replacement transaction spending the same inputs but paying a higher fee. Miners pick the higher-fee version. The original gets replaced and eventually drops out of mempools.

There’s also a method called CPFP — Child Pays for Parent. If you received an unconfirmed transaction, you can spend it in a new transaction with a very high fee. Miners who want that high fee must also confirm the parent transaction to collect it. You’re pulling a stuck transaction through by attaching a lucrative child to it.

Neither method works after confirmation. Before confirmation, they’re reliable tools — if your wallet supports them.

RBF reveals something broader: nodes and wallets sometimes follow slightly different rules. Which raises the question of what happens when larger parts of the network disagree on the rules entirely.

---

Bitcoin nodes running different software versions can coexist peacefully as long as the changes are backward-compatible — but when rule changes are incompatible, the network splits into two separate chains, each following its own rules independently.

Most Bitcoin upgrades are soft forks — changes that tighten or add rules in a backward-compatible way. Old nodes see new blocks as valid even if they don’t understand the new features. New nodes accept old blocks. The network stays together. SegWit was implemented this way — nodes that didn’t upgrade still functioned normally.

Hard forks are incompatible changes. A node following the old rules rejects blocks from new-rule nodes, and vice versa. If enough of the network goes hard fork, two chains emerge and continue independently — like Bitcoin Cash in 2017, when a faction insisted on larger blocks, split off, and created a separate currency that exists in parallel today.

The Bitcoin community learned through the block size wars that hard forks are dangerous and divisive. The ability to upgrade Bitcoin without forcing everyone to upgrade simultaneously is now a deeply held principle. Soft forks are preferred precisely because they don’t require unanimous agreement to move forward.

Understanding how the network handles software disagreements makes another question more concrete — what can governments actually do when they decide they don’t like any version of the software?

---

Governments can ban exchanges, restrict banking access, and block Bitcoin websites, but they cannot stop peer-to-peer Bitcoin transactions — the protocol runs over standard internet traffic and operates through VPNs, Tor, and even satellite connections.

Several countries have tried seriously. China banned exchanges in 2017, banned mining in 2021, and has repeatedly blocked Bitcoin-related services. Bitcoin usage in China continued through peer-to-peer trading, VPNs, and overseas exchanges throughout.

What governments can effectively control: on-ramps and off-ramps. If you need to convert local currency to Bitcoin through the formal financial system, that can be regulated, taxed, or shut down. Making Bitcoin inconvenient and legally risky through the traditional banking system is genuinely achievable and has been done.

What governments cannot control: the protocol itself. Bitcoin transactions are packets of data that look like regular internet traffic. They travel through VPNs and Tor. The Blockstream satellite broadcasts the entire Bitcoin blockchain from space. Bitcoin transactions have been sent via radio waves and SMS. The protocol doesn’t care about borders.

A sufficiently authoritarian government with full internet control can make Bitcoin extremely difficult to use practically. It cannot make it cryptographically impossible. The distinction matters.

Which takes the question to its logical extreme.

---

A complete global internet outage would temporarily halt Bitcoin transaction propagation, but the blockchain itself would survive intact on thousands of independent nodes, and the network would resume exactly where it left off once connectivity returned — no data lost, no history erased.

First, a true global internet outage is essentially impossible — the internet is itself a massively distributed network with no single off switch. Regional outages happen. Global does not.

But hypothetically: if global connectivity vanished, every node would freeze with its current copy of the blockchain. No new transactions could propagate. No new blocks could be found — mining requires network coordination to avoid wasted work. The blockchain would pause in place.

When connectivity returned, nodes would reconnect, share any blocks found during partial connectivity through satellite or radio links, resolve any chain divergence using the longest chain rule, and resume. The entire accumulated history would be intact. Nothing would be lost.

Bitcoin has actually survived significant regional internet disruptions — including connectivity failures in major mining regions — without losing data. The resilience isn’t theoretical. The architecture assumes unreliable connectivity, treats every node as potentially offline at any time, and routes around damage by design.

A global internet outage would inconvenience Bitcoin the way it would inconvenience everything else that runs on the internet. But it wouldn’t erase the ledger. The history lives in too many places simultaneously for that.

---

Related Deep Dive Threads

• Bitcoin Mining →
• Bitcoin Blockchain →
• Bitcoin Keys & Wallets →

Still curious? The transactions traveling through this network have their own internal structure — inputs, outputs, signatures, and a model of ownership that works nothing like a bank account. That rabbit hole is here.