Attack the network by having 51% of full nodes?

Question

Tell me if I'm missing something here:

If a government recognised that an adversery of theirs had a lot of bitcoin reserves, couldn't they just 'crash' the bitcoin network by spinning up enough full nodes to gain 51% of the network (not a hard thing to do for a big government)? They obviously couldn't produce fraudulent transactions or anything (talking about non-mining full nodes obviously). But they could just stop receiving/validating/appending new blocks. How would any particular node that blindly connects to one of these 51%'ers know not to interact with it? How would a given node 'know' who to query for chain data? Btw, I understand the government would have NO economic interest in doing this. They'd be purely knocking out someone else, but that's valuable enough surely.

Answer 1

The question has the appearance of mixing up two different types of attack

• The so-called "51% attack"
• The "eclipse attack"

The first is where an attacker controls >50% of the global mining hashpower, not 51% of nodes. This is an attack on every node in the network simultaneously.

The second is where an attacker controls all of the peers of a node. This is a targeted attack on a single node. Since a node might only have ten peers, this does not necessarily involve creating a huge number of dishonest nodes. It is also foiled if the attacked node has contact with only one honest node. That way it will become aware of a chain with more work than the attackers'.

Answer 2

If a government recognised that an adversery of theirs had a lot of bitcoin reserves, couldn't they just 'crash' the bitcoin network by spinning up enough full nodes to gain 51% of the network (not a hard thing to do for a big government)?

Technically most of us citizens are this adversary due to fiat economics.

How would any particular node that blindly connects to one of these 51%'ers know not to interact with it?

Likely because of what you said

they could just stop receiving/validating/appending new blocks

This causes people to switch node endpoints if they notice such a thing.

How would a given node 'know' who to query for chain data?

They would do it as usual, using network consensus rules to verify the longest chain.

Btw, I understand the government would have NO economic interest in doing this.

It's not that they wouldn't have interest in doing it, it's that the current state of the network renders their attempts absolutely obsolete. 50% attacks on Bitcoin are highly unlikely for a variety of reasons:

the only relevant nodes in terms of establishing network consensus (about new blocks) would be nodes that miners submit their proof of work to which are generally considered "mining pools". These pools are actually chosen BECAUSE of how secure and decentralized they are, when any of these pools act out of the interest of the miners (ie ordinals filling a full block on luxor pool) it causes them to fall out of the top pool ranks ie miners go somewhere else. Not to mention since these pools are so big, having a central actor or even a hand full of them control the whole thing is highly unlikely, if they did anything wrong miners would notice within the hour max and switch to another pool. If they were in a proof of stake system they would be frozen from withdrawing from the attackers pool. Considering this is the difference between pool maintainers making a good living and being bankrupt I would say the incentives are HIGH to keep it secure and decentralized.

I would really be entertained to see the counter arguments.