Drivechains — potential issues as they relate to Bitcoin
An overview of some possible issues with drivechains (a shorthand term for BIP300/BIP301) that may impact the Bitcoin network itself. These run counter to the argument that drivechains are ‘opt-in’ and that Bitcoin users that do not want to use them are unaffected by their activation.
I would like to thank Peter Todd for the discussions I have had with him on the topic and his valuable ideas and insights. Peter has published his analysis of drivechains here.
I have not focused on the mechanics of drivechains themselves, rather how they may impact Bitcoin.
Introduction
The ideas presented within this post are just high level ideas and are not accompanied by mathematical justifications. Indeed, the use of mathematical equations to “prove” that drivechains will not impact Bitcoin ignore the effects of human behaviour, unknown incentives, long term competitive advantages, seemingly irrational actions and the like. Experience tells us that it is not safe to assume you know and can model the behaviour of a large number of people. The desire to only activate Segwit as a hard fork but not as a soft fork springs to mind as an example of this, possibly the byproduct of a hidden competitive advantage. “Very unlikely to happen” is a phrase that has been used to argue against some forms of the issues raised in this post, but “very unlikely” may well still be too likely for most Bitcoiners who have wealth stored in the network.
Up front, I’m not saying that all these issues will happen and there may be many that can be satisfactorily answered by the proposers. My concern is that valid criticisms are not being acknowledged, understood, or addressed in enough detail and that the impression being generated is that the only criticism to drivechains is the misguided shouting of “No altcoins on Bitcoin!”. It’s certainly helpful for the promoters of drivechains that this is happening as it is easy to point at and label as the singular criticism and imply it is not one that should concern anyone. “Damn maxis!” etc.
If I say “drivechains” I am using shorthand for Paul’s sidechains proposal via BIP300/BIP301 (Hashrate escrows/blind merged mining). I also use 6 months as the peg out time period whereas in some cases I believe it can be configured to be less, for example I have seen 3 months stated in some documentation.
The post is a collection of notes, so pardon any poor grammar and lack of full references. More details can be provided on any item if needed.
I have asked Paul some of the questions contained within this post here but at the time of writing have not yet had responses. It’s also quite possible he has answered some of them in the past but “go listen to my 5 hour talk” (an actual suggestion during one of his Twitter Spaces sessions) is not a valid response in my opinion.
Possible ways it can impact Bitcoin
The potential for theft, remedied by Bitcoin UASF is part and parcel of the drivechain offering
This is perhaps the most common argument against the claim that drivechains are “opt-in” and do not affect Bitcoin users who do not use drivechains. The drivechain FAQ does address this but it makes assumptions up front about the incentives of miners and then builds everything else on top of that:
So the basis is that miners will act “just selfishly enough”… but what if they don’t stick to that, rather hopeful, assumption?!
Theft from one drivechain means cheap BTC for miners
What if a collection of miners decides they want to crash the 1:1 peg and devalue BTC on all drivechains by performing a theft from one drivechain? The breaking of trust in the 1:1 peg is realised, the users try to flee the drivechains and use the only viable route out en-masse, the swap providers, who do not want to take up drivechain BTC in 1:1 to mainchain BTC because of the risks now very much apparent. Miners could be the counterparty on swap services though — accepting (say) 3 DCBTC for every 1 BTC they swap. After they have accumulated many DCBTC they can then safely and “honestly” peg them out of the drivechain at a 1:1 rate, knowing that the people who control the peg out process that allows this are themselves. The panicked drivechain users are back in Bitcoin with less BTC than they had on the drivechain. There are many ways this panic could be created, not just via theft. Deliberate bugs introduced to drivechain code the miners created or sponsored, threats of splitting a drivechain’s blockchain etc.
If a pool does try a theft and promises to split the proceeds with their pool users — well, who would turn that down? This could occur if the drivechain is not generating enough fees and many other potential issues with the drivechain economics and codebase security itself. The theft itself can only be proven by those running a node on the drivechain, but how likely is that to happen?
Theft prevention via UASF
The comeback is to soft fork the theft transaction out on the Bitcoin chain, which impacts Bitcoin users.
What incentivizes miners to do anything now or in the future cannot be known, so best to keep it within the bounds currently set out in Bitcoin. Yes they like to make money but they may play a game where they take a hit now for profit later (breaking some of Paul’s models) or perhaps they have inside knowledge they use to prevent exit from a drivechain for later financial gain, or any number of unknowable situations that end up being valid in their eyes at the time but not in anyone else’s.
And if you don’t fancy the idea of miners doing this just out of greed and dishonesty then how about some government body in the US trying to stop miners from pegging out from a side chain of processing its transactions because they have been told not to.
The “theft being obvious and signalled for 6 months” argument assumes that there are people who are validating the sidechain itself other than the miners attempting the theft. How can you point to a proof of theft and tell others to look at it if you have no proof? Drivechain miners running nodes on the drivechain might block new attempts to connect to sync the blockchain if they don’t want people to have proof of a theft. It seems unlikely drivechain users will be encouraged to run drivechain nodes — after all this is touted as a way to onboard the masses.
And so a soft fork on Bitcoin is the last stand for these users to try and get their funds back. Bitcoin soft forks, or any type of chain split is not good for Bitcoin, it undermines the perception of stability and finality of transfers. Anything that puts a honeypot up and has a final defence of “soft fork Bitcoin” is very risky I would say and not “opt-in” at all. Drivechain users would be opting in other Bitcoin users to their own gamble.
Ruben SomSomsen commented (amongst others) on this topic 2 years ago at time of writing:
“…if the actual [Bitcoin] economic majority wants to prevent an invalid peg-out [theft transaction], this won’t lead to a permanent [Bitcoin] fork, but with a significant non-majority it could.” … “[Bitcoin] users are motivated to take an active stance against invalid peg-outs, and that’s concerning, as it turns Bitcoin consensus from passive and (relatively) certain to active and uncertain.”
The risks are known and yet still the “opt-in” line is used to try and ignore the fact that a change to the Bitcoin protocol via soft forks like this will place risk on Bitcoin itself.
Adds economic weight to one side of a contentious Bitcoin hard fork
Imagine the risk of a new version of the New York Agreement but this time the miners in support of it can hold drivechain Bitcoin hostage.
Those with BTC in drivechains would certainly voice their support of the miner side of the hard fork if miners said they would only release drivechain escrow on the Bitcoin fork they support and not the alternative. There would be ways around this on the other fork I guess, but this is far from an ideal power to place in the hands of miners.
A drivechain hardfork doubles the drivechain BTC supply
In the event of a hard fork on a drivechain, the BTC in the drivechain will effectively be doubled. The peg can only allow out as much as went in and so who determines which of the hard fork’s gets to peg out back to Bitcoin? This allows miners to effectively decide the fate of the forks on the drivechain and could possibly lead to ‘fork wars’ on the Bitcoin main chain as competing drivechain hard fork supporters try to soft fork out what they see as invalid withdrawals from the “lesser” drivechain chain.
What if there is a Bitcoin soft fork to remove drivechains after people have already sent BTC to them?
I haven’t thought too much about how this would come about from a technical standpoint but it doesn’t seem impossible. What happens to the funds in the escrow then? What if nobody was running a node on the drivechain apart from miners who now have access to all the funds and no longer the fees from the drivechain? There will possibly be no valid record of who deserves what allocation of BTC. This may result in campaigns to have funds released on Bitcoin and another potential Bitcoin soft fork.
The Bitcoin brand can be tainted by association
There are already people who think “Bitcoin was hacked” when an exchange is hacked, if there is theft from or on a drivechain then the association with Bitcoin could lead the wider public (the ones that drivechains claim to cater to for onboarding) to assume the position that Bitcoin itself is not secure.
Changes can add unknown future issues
Unintended consequences cannot be known upfront regardless of how much scrutiny a proposal has. We already have recently added changes to Bitcoin that we have yet to fully make use of so why rush this in so soon after? Segwit and Taproot were debated and thought through much more than drivechains but still brought about unintended consequences (ordinals for example). Why add another possible source of this so soon after?
Fee sniping becomes more likely
The likelihood of fee sniping to take the claimed “security budget solving” fees from drivechains (mining to re-organise so a miner can take a block where the fees were high enough to warrant the risks associated with lost work) increases, especially as the Bitcoin block subsidy decreases. Having the fees paid on an external chain affect Bitcoin’s own block production is far from ideal.
Impact on Bitcoin consensus and activation processes
There are also risks to Bitcoin’s change process, not just the blockchain itself.
Why is this now being presented as urgent with talk of miner activation incoming? Why are they not launching a bip8 signal/flag process or something? Why are they not trying to build consensus first? If they get miners to activate without user consensus it sets a dangerous precedent for future soft forks. “Fire first and deal with any backlash (URSF)” is not a sound way to make changes to Bitcoin.
David Bailey claims that he has been talking to miners and they say (to put it more politely) that they will activate in a hasty way, regardless of the general sentiment of Bitcoin users.
Discussions have been ongoing for many years, it is true, but fundamental issues with incentive skews on Bitcoin have not been addressed. The lack of support is not just because of people not liking altcoins, which is what some proponents have implied. The idea of a decentralised, trustless, two way peg is great, perhaps this is just not the way to do it and the right way way may not present itself for many years, if at all. So be it, there is no do-over in Bitcoin.
Other common or interesting points
Mike Tidwell has a thread where he collects interesting takes on potential issues with drivechains. The thread is active and being added to so refer to the source for an up to date list. Some of the points raised by others there have already been covered in this post but it is worth noting them for a wider view on how they may come about or what impact they may have.
These cover items Mike has labelled: “No one wants it”, “Elsewhere first”, “Stratum v2 first”, “The unpeggining”, “Insufficient fees won’t secure”, “Adoption is dangerous”, “Key-signing federation is superior”, “The BMM fallacy”, “Doesn’t kill alts”.
Paul has responded to one or two at the time of writing.
Decentralised, trustless 2WP — fact or fiction?
Trusted custodians
The proposal’s own website suggests that, due to the limited peg out capacity, the users would likely opt for a custodial swap service or exchange use:
If you want a Monero like drivechain for a transaction you wish to make, why not just use Monero right now via an existing swap service and swap back to Bitcoin after you have used the feature you wanted?
Liquidity Swap services like the ones mentioned in the above image will need to rebalance the sides of their swaps and so how many of the 20,000 outputs will be available for normal people to use? It may also be financially beneficial for such services to pay the fees to take up all those outputs for themselves, leaving drivechain users with little options beyond paying service providers to get out. As for atomic swaps, finding someone who happens to want in at the same time and for the same amount will not be easy, especially if there is a “panic” flow of users trying to leave all at once. So if you don’t like the idea of Bitgo being the custodian of WBTC then why would you like drivechains which, baring some person to person coincidence of wants atomic swaps, will likely end up relying on such custodians for drivechains?
The peg will not remain priced 1:1
It is very easy to think up multiple scenarios where the lack of exit capacity (20,000 outputs every 6 months for example) and possible panic scenarios causes a swap service to re-price the drivechain BTC lower than mainchain BTC.
Creating drivechains is unlikely to be open to everyone as advertised
As “zndtoshi” points out (refer to Tweet for further details):
“BIP300 requires 90% of the miners to vote in favor for a new sidechain to be enabled. This also means that 10% can veto a new sidechain proposal.
Drivechain proponents assume that miners will behave in a good way, while in reality they have no guarantee for this.
It’s possible…
..that miners could take advantage of this power to veto and expect benefits.”
Overselling, misrepresenting, or not correcting the mistakes of others
The way in which some issues raised have been ignored or “hand waved” away has caused me to dig deeper into them. The lack of adversarial thinking was what led me to create this post. The main website for the drivechains proposal is of concern as it does not acknowledge issues with enough consideration for Bitcoin which, it seems to me, is perhaps treated as a vehicle to deliver the prosper’s own prediction market project.
Bitcoin (the asset) does not make a drivechain secure or less risky just because it is pegged into it
Just because a drivechain uses pegged bitcoin on it doesn’t remove bugs, potential scams or malicious actors not constrained by the network effect accrued by Bitcoin over the years. The risk is that the peg to Bitcoin may well enable the creator of any drivechain to sell it as being “safe because it’s pegged to Bitcoin, it’s a Bitcoin layer 2, not an alt coin”.
It absolutely won’t make “Bitcoin twitter” an “everyone is happy”, contention-free environment. Just because the “we want more privacy” people have a drivechain with Monero-like properties doesn’t mean that’s them done and happy, they’d then likely point to their drivechain and say “see, works fine, let’s add it to the main chain, then we can all benefit and I can stop having to peg out when I want to pay someone in a normal bitcoin address”. But a drivechain testnet is not the same thing as the bitcoin main chain. There is zero chance this will end disputes and may even provide “evidence” that sparks more.
Implying double spends are the same thing as drivechain escrow theft
There is a big difference between a 51% attack double spend on Bitcoin and a theft from a drivechain’s escrow fund. The two are presented as the same by drivechain proponents in order to dismiss concerns over the magnitude of potential escrow theft along the lines of “well miners can already steal using a 51% attack in Bitcoin so the same thing that stops them doing that will stop them stealing from the drivechain”.
Apparent lack of awareness of pre-existing issues raised
Worrying to me is that the issues Shinobi raised on the 1st sept Spaces discussion were not understood by Paul despite them being repeated 5–6 times. This was not a new issue that was being raised either (fees from drivechains can centralise mining). Michael Tidwell, who stepped in to act as a ‘moderator’ in this part, understood the point being made and explained it to Paul in different terms. Paul still did not acknowledge the issue, instead changing the context or saying he didn’t see what the issue was. Not getting it or, less charitably, pretending not to get it in order to avoid addressing it, are both worrying in equal measure!
Fear/greed based or dishonest framing
https://twitter.com/LayerTwoLabs/status/1664032295056191488
The above tweet from the proposer’s LayerTwoLabs account mentions:
“100% of the value of many utility tokens such as monero, ethereum, and others may begin to be absorbed by Bitcoin…”
“Aside from huge price appreciation of #Bitcoin…”
“solving the security budget problem…”
“more on-chain onboarding […] (instead of scaling via trusted/custodial services)…”
“Drivechain relies on the same or similar assumptions on which Bitcoin was designed”
“Bitcoin needs Drivechain or similar, or there may eventually be a flippening, loss of network effects, insufficient fee-based security, and other major problems for Bitcoin.”
All I can picture is a fragmented market where BTC moves out of the main market to disconnected islands of lesser liquidity. I can’t pay for some socks in bitcoin because I moved half my bitcoin to a Monero-like drivechain and the rest to an EVM based drivechain.
“Proving” incentives using spreadsheets
The drivechain FAQs are vague and assume everything can be represented by a figure in a cell in a spreadsheet. Vague answers, blunt assurances, or relying on spreadsheet maths is not very convincing to me. Human emotions, incentives and drivers cannot be represented by a number in a spreadsheet.
It is certainly not urgent!
The proposers and supporters act as if the activation of BIPS 300 and 301 by soft fork is urgent. This is perhaps due to recent funding and the associated expectations of delivery that brings about.
Some of the more prominent supporters are even displaying tweets showing that they are canvassing for miner support to activate via MASF before there is clear consensus among users.
If drivechains are so great it will be worth taking our time over and worth them clearly addressing issues and being up front and honest about any potential drawbacks. If not, the proposal will receive hostile responses, regardless of its own merits.
Trying to get bitcoiners to rally behind it to “stop Ethereum flippening Bitcoin” or that with drivechains we’d be at $15 million a coin by now is playing to the lowest human emotions of fear and greed and is not winning the author of this content over. In fact it has had the opposite effect and has made me question why such tactics are needed if the idea is as good as is claimed.
Footnote
I wrote most of the above in early September so some references might now be replaceable with more relevant ones and I may have missed linking to more recent thoughts on the above matters by others in the space.
