May we have 2 independent networks?

[PawGo]

Recently I have found an information about problems with fiber cable which was damaged (probably intentionally) and causes some networking issues:
https://trust.zscaler.com/zscloud.net/posts/12256

It made me think - is it possible that bitcoin nodes network would be split? I know that many connects to many etc, but what if "2 bubbles", similar size, enough to be fully functional, would be created? Imagine that for example South&North Americas lost connection to rest of network (no oceanic cables, no satelites etc etc).
Is it possible to detect that you are in a separated bubble? I assume we would have 2 chains, which would make it impossible to merge.
From the point of single node it would only mean that you lost some connections but you are able to quickly find replacements. At the end now, when you have your node and you are connected to 10 other nodes, you do not know if you are connected to Asia and Australia or to your neighbors on the next street - there is no criteria for nodes selection, is there?

[1714319298]

1714319298

[1714319298]

1714319298

[1714319298]

1714319298

[o_e_l_e_o]

Is it possible to detect that you are in a separated bubble?

If the network was split fairly evenly, both sides would immediate notice block times roughly doubling due to hashrate halving until we hit the next retarget. Outside of this, though, I'm certain that everyone would immediately notice the internet splitting in two, since you would not be able to communicate with many of your contacts and you would lose access to huge chunks of content.

I assume we would have 2 chains, which would make it impossible to merge.

Provided neither chain rolled out any update or fork which would make it incompatible with the other chain, then when the connection was reestablished whichever chain had the most proof of work would win and the other chain would simply be abandoned. I would assume in such a doomsday event though that at least one side would deliberately take steps to prevent this from happening so as not to risk a massive reorg and loss of hundreds or even thousands of blocks worth of block rewards.

there is no criteria for nodes selection, is there?

In terms of geography then no, not as far as I am aware, and with nodes being run over Tor/VPN/VPS then IP addresses are not reliable indicators anyway.

[ABCbits]

From the point of single node it would only mean that you lost some connections but you are able to quickly find replacements. At the end now, when you have your node and you are connected to 10 other nodes, you do not know if you are connected to Asia and Australia or to your neighbors on the next street

Each country have different IP address allocation, so you could make educated guess based on other node IP.

there is no criteria for nodes selection, is there?

It depends on which full node implementation you use. IIRC Bitcoin Core avoid make connection to multiple nodes with same IP block.

[PawGo]

From the point of single node it would only mean that you lost some connections but you are able to quickly find replacements. At the end now, when you have your node and you are connected to 10 other nodes, you do not know if you are connected to Asia and Australia or to your neighbors on the next street

Each country have different IP address allocation, so you could make educated guess based on other node IP.

ip to location may be helpful indeed, for not-tor connections. It would be interesting to have something like a quota for different continents/countries.

there is no criteria for nodes selection, is there?

It depends on which full node implementation you use. IIRC Bitcoin Core avoid make connection to multiple nodes with same IP block.

Interesting, I did not know that. There may be speed test, but it would favor close nodes, so it is not the best solution.

So, as I understand if let's say one country with power big enough to mine blocks is disconnected, they build their own chain, but their work will be lost after reconnection.

[BlackHatCoiner]

Possible duplicate of: What will happen to the Bitcoin network if Russia blocks access to the internet?

there is no criteria for nodes selection, is there?

There are various methods for node selection, but there are no topographical parameters taken into consideration.

ip to location may be helpful indeed, for not-tor connections. It would be interesting to have something like a quota for different continents/countries.

You can check this here: https://bitnodes.io/

[DaveF]

In the scientific view of things it is an interesting scenario to discuss. In reality with the way the the internet is setup and the way things work it's just about impossible.

There are so many redundant paths to places and contiguous loops of fiber and BGP is really robust enough to figure out a way to get data from A to B. It *will* be painfully slow and for BTC there would probably be a bunch of 1 or 2 block reorgs as nodes get the new tip of the blockchain. But, since international banking and finance and shipping and just about everything else would have imploded it would probably not matter.

Someone brought this up a couple of weeks ago too. Can't find the post at the moment.

-Dave

[jackg]

I'm not too sure those maps of where Internet cables are located are actually that accurate - there's probably more cables connecting places like Europe and North America for example that aren't included, perhaps some connecting places like Australia, Indonesia, India, China and the US too so it'd take a lot to try to split the connections completely.

If it did happen, most online systems would end up failing anyway so we'd have that to deal with. There would be other ways of transmitting data, such as satellites, which might be repurposed for such things although could also be slow (unless well planned and implemented).

As noted above, block speeds would be very noticeable from the onset, if Asia has a lot more miners than America then the one with the least power behind it would become really slow and the longest chain would win out over the shortest chain when a reconnection happened. If the reconnection happens after 12 hours passes, I'm unsure how a double spend attack could actually be stopped (without the chain being considered entirely insecure and companies stopping dealings in it for a while - since you could hop on a flight or send your private key to a contact on the other side of the disconnect just after sending your transaction and selling your funds).

Information for critical infrastructure could also be flown across if it needed to be and I don't think those flights would be too slow with military involvement (less than 2 hours).

[spectre71]

The internet is so robust that this doomsday event is darn near impossible. Satellites would need to taken down as well.  Someone mentioned BGP, yes in a nutshell here it's a self healing protocol to automatically reroute after failures.

[o_e_l_e_o]

As noted above, block speeds would be very noticeable from the onset, if Asia has a lot more miners than America then the one with the least power behind it would become really slow and the longest chain would win out over the shortest chain when a reconnection happened.

If the split was very one sided, e.g. 95%/5%, the chain with 5% of the hashrate would be facing block times of 3 hours or more and up to 9 months before the next retargeting. In such a case, their chain would become completely unusable unless they quickly implemented some new retargeting mechanism. Worth noting that in such a case it would be the chain with the most proof of work which would win in the case of a reconnection, which is not necessarily the longest chain. If the chain with 95% continued to mine at the current difficulty, but the chain with 5% forked to drop the difficulty drastically, then the 5% chain could end up being longer, but the 95% chain would win due to having accumulated more work in the production of its chain.

[stompix]

If the reconnection happens after 12 hours passes, I'm unsure how a double spend attack could actually be stopped (without the chain being considered entirely insecure and companies stopping dealings in it for a while - since you could hop on a flight or send your private key to a contact on the other side of the disconnect just after sending your transaction and selling your funds).

Rather than 12 hours, I wonder what would happen if this will be just a one-two hours event and it will one that will split the networks in no more than a 2:1 hashrate ratio, so in one hour even the smaller chain will be able to mine quite a few blocks. By the time people realize it has been a split and they will try to do something more blocks will be added, more headaches, and so on and on.

That aside, a complete cut at global levels is nearly impossible, a country-level one is definitely in the books but it will have way lower implications.

[o_e_l_e_o]

Rather than 12 hours, I wonder what would happen if this will be just a one-two hours event and it will one that will split the networks in no more than a 2:1 hashrate ratio, so in one hour even the smaller chain will be able to mine quite a few blocks.

Then when communication is reestablished, the chain with less proof of work (which would almost certainly be the "1" chain in 2:1 split) would simply be abandoned as all their nodes reorg to accept the chain with more work. There will not be enough time in 1-2 hours to do anything else.

With such a short split, there might not be that much disruption. Assuming that no one has access to both chains (otherwise there wouldn't be a split at all), then every transaction on the minority chain will still be valid and could then be mined on the main chain after the reorg. There would be no transactions invalidated due to block rewards being invalidated, since we would be well below the 100 block lock up period.

[Ucy]

Issue like this has been looked at in the past, solutions were developed, one of which is based on sidechain/multi-chain (PoW) Networks. Once a chain in the multi-chain Networks lost connection to the rest of the Networks or is detached from the rest of the world, it can continue to function in its local environment as a single blockchain with its own token and still have the ability to exchange existing bitcoins in the local blockchain with local peers. I think the price of the local bitcoin will vary from the rest of global market depending on how much the split networks value it.

The split network can rejoin the rest of the global multi-chain networks (when connection is re-established)  and continues to exchange its bitcoins with the networks, or it could exist as a separate network within the multi-chain networks with its own token (or colored coin).

In such multi-chain networks, it's best to store your coin in the local blockchain you control and can easily run as node. If you detach your blockhain network from the rest of the multi chain networks, your chain/node can continue to function and exchange token/bitcoin with your local peers.




If that happens to this current Bitcoin Network, Multiple Crypto networks would eventually be created.  But I think there will be alternatives means for people to exchange and verify bitcoins with established international Bitcoin Networks. For example radio frequencies, via sms etc

[NotATether]

For the record, it is perfectly possible for two isolated networks to form (assuming satellite internet is non-existant), but then when the two networks regain communication with each other, one of the states will be wiped out - the one with a shorter chain tip. It is not possible for networks to "diverge to the point of no return" because they can always drop their divergent chain for the longest one.

[ABCbits]

It depends on which full node implementation you use. IIRC Bitcoin Core avoid make connection to multiple nodes with same IP block.

Interesting, I did not know that. There may be speed test, but it would favor close nodes, so it is not the best solution.

Based on commentary[1] i found, it looks looks like what Bitcoin Core do is bucket/split the node by /16 IP group.

[1] https://github.com/bitcoin/bitcoin/issues/16599

[o_e_l_e_o]

It is not possible for networks to "diverge to the point of no return" because they can always drop their divergent chain for the longest one.

Sure, but the issue is that the chains might have diverged so much that to abandon the chain with less proof of work would mean reversing weeks or months or payments, wiping thousands of bitcoin completely out of existence due to lost block rewards, and completely invalidating every transaction which has one of these lost block rewards anywhere in its history. The combined chaos of all that happening might mean that the people using the minority chain deliberately make changes to their protocol to prevent their nodes from abandoning their chain in the event of communications being reestablished, leaving us with two networks running side by side.

Based on commentary[1] i found, it looks looks like what Bitcoin Core do is bucket/split the node by /16 IP group.

It splits it by /16 for IPv4 and /32 for IPv6, and will try not to pick nodes from the same bucket. But as I said above, IP addresses aren't a super reliable indicator of geography, and even if they were, it would still be easily possible to pick 10 nodes from 10 different buckets all within the same country.

[tam31]

I don't think it’s possible, is it? I mean if we have two different chains then it's two different chains, we cannot have two different networks and then call it the same thing. Meaning we could have unlimited number of networks not just 2, but they won't be the same thing and this is where the problem starts.

Independent networks would be totally different from each other and that’s a bit of a deal, and can't be used interchangeably when something happens to one of them. But we shouldn't be worried, bitcoin is too strong to have these kinds of issues, it's strong and even if there is something wrong, we move on and just don't worry about anything else.

[DannyHamilton]

For those that happen to be curious whether Satoshi ever eve thought about this possibility:

It's hard to imagine the Internet getting segmented airtight.  It would have to be a country deliberately and totally cutting itself off from the rest of the world.

Any node with access to both sides would automatically flow the block chain over, such as someone getting around the blockade with a dial-up modem or sat-phone.  It would only take one node to do it.  Anyone who wants to keep doing business would be motivated.

If the network is segmented and then recombines, any transactions in the shorter fork that were not also in the longer fork are released into the transaction pool again and are eligible to get into future blocks.  Their number of confirmations would start over.

If anyone took advantage of the segmentation to double-spend, such that there are different spends of the same money on each side, then the double-spends in the shorter fork lose out and go to 0/unconfirmed and stay that way.

It wouldn't be easy to take advantage of the segmentation to double-spend.  If it's impossible to communicate from one side to the other, how are you going to put a spend on each side?  If there is a way, then probably someone else is also using it to flow the block chain over.

You would usually know whether you're in the smaller segment.  For example, if your country cuts itself off from the rest of the world, the rest of the world is the larger segment.  If you're in the smaller segment, you should assume nothing is confirmed.

As o_e_l_e_o points out, if the segmentation lasts long enough for miners to start spending block rewards from blocks that are mined during the segmentation, then there is potential for those transactions to become invalid when the chains re-combine.  If those on the smaller segment are foolish enough to trust the network while on the smaller segment, then there is incentive for them to create forked software that maintains their chain separate from the main Bitcoin chain.  Whether that forked chain actually maintains any useful value in the world will probably depend a lot on the world economic situation at that time.

[DaveF]

It is not possible for networks to "diverge to the point of no return" because they can always drop their divergent chain for the longest one.

Sure, but the issue is that the chains might have diverged so much that to abandon the chain with less proof of work would mean reversing weeks or months or payments, wiping thousands of bitcoin completely out of existence due to lost block rewards, and completely invalidating every transaction which has one of these lost block rewards anywhere in its history. The combined chaos of all that happening might mean that the people using the minority chain deliberately make changes to their protocol to prevent their nodes from abandoning their chain in the event of communications being reestablished, leaving us with two networks running side by side.

As I said before if the global internet suffered such an outage for such a long period of time we are going to be living in a MadMax / World War Z kind of scenario so at least for the few of us that are still alive it's not going to matter much.

Interesting enough, in the event of a global internet outage big cities and suburbia will probably have a much better time of it then the middle of nowhere. With most of the large data centers in populated areas it would be easier for them to get access back to the people close to them, then someone living in the middle of a farm 200 miles from nowhere.

-Dave

[o_e_l_e_o]

I mean if we have two different chains then it's two different chains, we cannot have two different networks and then call it the same thing. Meaning we could have unlimited number of networks not just 2, but they won't be the same thing and this is where the problem starts.

In the event of an internet split, then it is entire possible to have two different chains which are both called bitcoin. They would both be running the exact same code, and as far as each node is concerned, would be following the consensus rules by following the chain with the most proof of work (since they are unaware of the alternative chain). And the same thing could happen 3, 4, 5, or more times. Each independent network would still be running bitcoin as far as they are concerned. But as discussed above, we already have a mechanism in place to resolve such a split when the networks rejoin.

Interesting enough, in the event of a global internet outage big cities and suburbia will probably have a much better time of it then the middle of nowhere. With most of the large data centers in populated areas it would be easier for them to get access back to the people close to them, then someone living in the middle of a farm 200 miles from nowhere.

True, but in a global internet outage supply chains collapse almost immediately, so I'd much rather be living 200 miles from anywhere with my own back up generator, clean rainwater collection system, and land for growing food, than in the middle of a heavily populated city which would quickly break down in to constant riots, looting, and violence, over the quickly diminishing supply of food and other essentials.