"Node war" possible with spoofed version strings? - page 2.

achow101

staff

Activity: 3458

Merit: 6793

Just writing some code

Quote from: tsoPANos on August 20, 2015, 03:48:04 PM

Quote from: YarkoL on August 20, 2015, 03:39:27 PM

I must be missing something, but I thought the fork depended on blocks
mined, not on number of nodes.

So maybe the anti-XT party could spend some electricity to
mine "fake" XT blocks in order to produce illusion of 75% goal
reached. But would that really accomplish anything? Undecided miners might
still jump on the bandwagon.

Well that will not cause premature >1mb blocks, as they are scheduled no earlier than 11 Jan 2016.

However it could cause the fork to occur on 11 jan 2016 without consensus. If for example 50% of the miners were XT, 25% NotBitcoinXT and the rest Core, then the fork could happen and spawn two chains with equal hash power. This could be detrimental to Bitcoin has we have previously established that such a fork is not a good thing.

tsoPANos

hero member

Activity: 602

Merit: 500

In math we trust.

Quote from: YarkoL on August 20, 2015, 03:39:27 PM

I must be missing something, but I thought the fork depended on blocks
mined, not on number of nodes.

So maybe the anti-XT party could spend some electricity to
mine "fake" XT blocks in order to produce illusion of 75% goal
reached. But would that really accomplish anything? Undecided miners might
still jump on the bandwagon.

Well that will not cause premature >1mb blocks, as they are scheduled no earlier than 11 Jan 2016.

YarkoL

legendary

Activity: 996

Merit: 1013

I must be missing something, but I thought the fork depended on blocks
mined, not on number of nodes.

So maybe the anti-XT party could spend some electricity to
mine "fake" XT blocks in order to produce illusion of 75% goal
reached. But would that really accomplish anything? Undecided miners might
still jump on the bandwagon.

mezzomix

legendary

Activity: 2702

Merit: 1261

Quote from: Cryddit on August 19, 2015, 07:03:06 PM

An attempt to sabotage the process is a sure sign that the anti-BitcoinXT people fully understand that they would lose in a fair consensus decision.

No. It's a clear answer to the way some people try to enforce a change of the consensus rules. Everybody is free to to choose his own version string and consensus rules. BXT shows that communication with the people to reach almost 100% consensus before doing a change is no longer necessary.

In the future we now might see more changes of the consensus rules by small groups of people trying force their ideas to be accepted by the majority. Live with it!

Cryddit

legendary

Activity: 924

Merit: 1132

An attempt to sabotage the process is a sure sign that the anti-BitcoinXT people fully understand that they would lose in a fair consensus decision.

And, to me, that means the consensus decision (the real one) has already been made and now there's nothing left for them but this kind of screaming and FUD tactics and trying to prevent people from accurately seeing what the decision is.

mmortal03

legendary

Activity: 1762

Merit: 1011

Excellent. So, the Core devs didn't release it, but someone did.

DannyHamilton

legendary

Activity: 3472

Merit: 4801

Quote from: mmortal03 on August 17, 2015, 05:54:26 PM

One respondent to that fake Satoshi message from the other day brought up the concept of a potential "node war" by way of spoofed version strings.
- snip -
I doubt Core would actually do something like this, but, if the version string actually has no bearing on the actual code running on a said node
- snip -

https://bitcointalksearch.org/topic/not-bitcoin-xt-1154520

Quote from: chek2fire on August 16, 2015, 07:42:25 PM

This is a special fork for those who do not agree with the blocksize scheduled increase as proposed by Gavin and Mike in their divisive altcoin fork, "Bitcoin XT".
- snip -
This version is indistinguishable from Bitcoin XT 0.11A except that it will not actually hard fork to BIP101, yet appears on the p2p network as Bitcoin XT 0.11A replete with features, yet at a consensus level behaves just like Bitcoin Core 0.11. If it is used to mine, it will produce XT block versions without actually supporting >1MB blocks.

Running this version and/or mining with XT block versions will make it impossible for the Bitcoin XT network to detect the correct switchover and cause a premature fork of anyone foolish enough to support BIP101 without wide consensus from the technical community.

It prevents correct detection of Bitcoin XT adoption in the wild since usage will be known to have been tampered with and thus all statistical data gathered by getnodes can only be considered unreliable.

Carlton Banks

legendary

Activity: 3430

Merit: 3080

Quote from: mmortal03 on August 17, 2015, 05:54:26 PM

Could there be, to counteract this, a more verifiable way developed to know that the version string being displayed *actually* matches the code that is being run underneath all the nodes?

It would likely become more cat-and-mouse dynamics. It seems like something close to 99% consensus is what you really need to implement a successful hard fork, and there's no substitute for that.

mmortal03

legendary

Activity: 1762

Merit: 1011

One respondent to that fake Satoshi message from the other day brought up the concept of a potential "node war" by way of spoofed version strings. He said,

Quote

Core could appropriate the version string of XT, making it impossible to know how much they are progressing and a losing bet to actually execute the fork.

Source: http://lists.linuxfoundation.org/pipermail/bitcoin-dev/2015-August/010252.html

I doubt Core would actually do something like this, but, if the version string actually has no bearing on the actual code running on a said node, then is this way of voting by node version string really a foolproof technical solution to achieving consensus? Can this sort of spoofing be a reasonable attack vector? Could there be, to counteract this, a more verifiable way developed to know that the version string being displayed *actually* matches the code that is being run underneath all the nodes?

Topic: "Node war" possible with spoofed version strings? - page 2. (Read 3541 times)