Is it possable to create proof of stake sidechains?

[Anon136]

It would be very nice to have a functioning proof of stake sidechain just in case proof of work ever failed us. Perhaps some malicious actor produces asics in bulk at a nano-meter scale above the current best. Maybe a manufacturer creating cutting edge devices decides to sell them all to some shadowy alphabet soup agency. Maybe pools that appear on the surface to be distinct entities are actually the same entity. Its not hard to imagine proof of work dooms day scenarios.

The paper talks about a lot of interesting ways that bitcoin could be experimented with, but it doesnt seem to make mention of experimenting with side chains with alternate consensus mechanisms. I wondering if this is because it is assumed that all sidechains would necessarily need to rely on bitcoin miners for security. I mean while that resource is there and basically a public good for sidechains it definitely makes sense to make use of it, but it would be very nice to have alternate consensus mechanisms in sidechains not so much for immediate use, but as a contingency.

As far as I can tell from reading the whitepaper the basic idea is to "lock" your bitcoins in a special transaction on the bitcoin block chain, and then if you wish to ever remove the coins from a side chain you dont actually need to make any sort of transaction for that on the side chain, you just undo that lock on the bitcoins on the bitcoin block chain, and then participants in the side chain will consider the coins that used to be locked to be an invalid input to a transaction on their chain. If I am understanding this correctly than it should be perfectly possible to create alternative consensus mechanisms on sidechains.

Sorry in advance if this was already explained in the white paper and it just went over my head and thanks in advance to the brainiacs who respond to my senseless rambling.

[1714975354]

1714975354

[DeathAndTaxes]

Sidechains are dependent on the security of the main chain.  So if proof of work fails then all associated sidechains fail as well.  Consensus on the sidechain can be done however you wish.  It could use proof of stake, some as of yet undeveloped concept, or even a centralized validation server (i.e. paypal bitcoin sidechain).

[andytoshi]

For a 2-way pegged sidechain to work, the parent chain must be able to understand proofs of consensus in the child chain. So you could not directly have a proof of stake sidechain hanging off of Bitcoin; Bitcoin does not understand stake proofs, nor are there any plans for it to understand stake proofs, nor can I imagine there ever being such plans, because my view of the Bitcoin community is that we'd never get consensus on this.

This is fine. Whatever mechanism for sidechains does exist, we can create "adaptor" sidechains which are children of Bitcoin and parents of other chains. They'd use a proof-of-work that Bitcoin understands, but their transaction script would understand proofs of [insert consensus mechanism here] to allow them to be parents. Then to move Bitcoins onto the pos-chain and back you'd have to first go through the adaptor sidechain. (Atomic swaps, of course, would not need to use the adaptor chain.)

Having said all that, it's really not clear that proof-of-stake chains can be sidechains at all. It appears to be impossible to form consensus from proof of stake without external trusted parties, but a sidechain peg necessarily has to be mechanical and unable to access external trusted parties. Can you clarify how you imagine this working?

[DeathAndTaxes]

Bitcoin doesn't also understand SPV proofs for a trustless 2-way peg either and there are no plans to change that via a fork at this time. The only thing that is currently possible would be a federated trust to secure the peg and that would work just as well is the sidechain used PoS.

As far as PoS being unable to form a consensus well that is a larger issue about PoS.  Not everyone shares that opinion.  A sidechain using PoS wouldn't be in any worse of a situation than a primary chain using PoS.

[Anon136]

thanks guys. appreciate you taking the time to address my question. 

[Anon136]

For a 2-way pegged sidechain to work, the parent chain must be able to understand proofs of consensus in the child chain. So you could not directly have a proof of stake sidechain hanging off of Bitcoin; Bitcoin does not understand stake proofs, nor are there any plans for it to understand stake proofs, nor can I imagine there ever being such plans, because my view of the Bitcoin community is that we'd never get consensus on this.

This is fine. Whatever mechanism for sidechains does exist, we can create "adaptor" sidechains which are children of Bitcoin and parents of other chains. They'd use a proof-of-work that Bitcoin understands, but their transaction script would understand proofs of [insert consensus mechanism here] to allow them to be parents. Then to move Bitcoins onto the pos-chain and back you'd have to first go through the adaptor sidechain. (Atomic swaps, of course, would not need to use the adaptor chain.)

Having said all that, it's really not clear that proof-of-stake chains can be sidechains at all. It appears to be impossible to form consensus from proof of stake without external trusted parties, but a sidechain peg necessarily has to be mechanical and unable to access external trusted parties. Can you clarify how you imagine this working?

I don't know about side chains specifically but if you want to talk about whether proof of stake can work more generally the answer is yes it definitely can. BCNext and cunicula collaborated to figure it out and their solution was genius. I just assume it would work the same way that nxt works. Do you want to talk about how nxt works under the hood? It is a bit off topic but i dont mind at all.

[andytoshi]

As far as PoS being unable to form a consensus well that is a larger issue about PoS.  Not everyone shares that opinion.  A sidechain using PoS wouldn't be in any worse of a situation than a primary chain using PoS.

It would be in a worse situation because the parent chain's concept of time is even more decoupled from the sidechain's concept time than clock time is. It has no ability to appeal to human trust networks. And actions on the parent chain are irreversible; there can be no "flag days" or concerted efforts to change its view of the sidechain's history, unless it is willing to break its own consensus on behalf of the sidechain.

I don't know about side chains specifically but if you want to talk about whether proof of stake can work more generally the answer is yes it definitely can. BCNext and cunicula collaborated to figure it out and their solution was genius. I just assume it would work the same way that nxt works. Do you want to talk about how nxt works under the hood? It is a bit off topic but i dont mind at all.

Unless you are going to explain my costless simulation argument, show where it is incorrect, and show how this mistake can be exploited to produce a counterexample to my argument, the answer is absolutely not. (And to be clear: of the dozens (hundreds?) of altcoins which have claimed I am wrong with no justification, you would be the first to do so.) I have been paid exactly $0 for engaging in these sorts of arguments and am apparently always the only participant to be bothered by correctness. I have no more energy for it.

[Anon136]

the answer is absolutely not.

k. well once again thanks for taking the time to reply to my OP.

[samson]

the answer is absolutely not.

k. well once again thanks for taking the time to reply to my OP.

There you have it : it's never ever going to happen.

Zero flexibility

[Anon136]

the answer is absolutely not.

k. well once again thanks for taking the time to reply to my OP.

There you have it : it's never ever going to happen.

Zero flexibility

Honestly. I probably am not capable of critiquing his paper in any meaningful way. That's the only reason I declined. I simply am not as knowledgeable or intelligent as a lot of people here (especially in the development and technical discussion section). One thing I do know is how nxt works because I put a lot of work into it and so I am capable of discussing the problem from that particular angle. He wasn't interested in that though, so there isnt much i can offer here.

[Gabriel Eiger]

Bitcredit/CRE implements what you are looking for in a "soft" sidechain where bitcoins can be claimed continually from the Bitcoin blockchain and the Bitcredit internal mining algorithm is combined pow/pos, heavily weighted towards pos.

See the section on sidechaining in the Q&A in the announcement for further explanation.

https://bitcointalk.org/index.php?topic=995970.0

[Gabriel Eiger]

Bitcredit/CRE implements what you are looking for in a "soft" sidechain where bitcoins can be claimed continually from the Bitcoin blockchain and the Bitcredit internal mining algorithm is combined pow/pos, heavily weighted towards pos.

See the section on sidechaining in the Q&A in the announcement for further explanation.

https://bitcointalk.org/index.php?topic=995970.0

Taken from the Q&A section:

Is this a sidechain?
- Yes and no. It is NOT a sidechain implemented through a two-way-peg that removes coins from the Bitcoin network with the possibility to return them later. It is however, a working implementation of a one-way claiming process that taints the bitcoins that already have been claimed. This is a process that in practical terms makes Bitcredit a side chain. Sidechains that actually can move coins back and forth between chains can not be implemented in Bitcoin without major updates to the Bitcoin protocol. It can also be discussed if this actually is what one wants to achieve in a sidechain implentation, instead of giving Bitcoin users claiming possibilities in a new currency that has different properties, thereby enabling a safer transition to a new system. The Bitcredit implementation can be considered this type of "soft" sidechain implementation, since it does not destroy bitcoins. It is, to be more precise, a variant of what is described under section 3.3. Assymetric two-way-peg in the blockstream white paper on side chains. http://www.blockstream.com/sidechains.pdf.

[Anon136]

thanks for that info

[spartacusrex]

Couple of points / ideas I've been mulling over..

If you had Bitcoin's POW as the main-chain, and then a POS sidechain, could this not remove the 'long range attacks', as the checkpoints required to fix the issue would be stored on the POW chain ? So there would be no 'weak subjectivity' required  ? You can always tell what a valid checkpoint in the valid POS chain is, by checking the un-fakeable POW chain.

This leaves only the short range attacks, and as andytoshi says, there is always the 'costless simulation' that can be used to manufacture many-many chains until you find one you like.

There seem to be 2 counter arguments to that currently.

1) Vitalik's ideas about a punitive POS chain (Slasher) that stops this sort of nonsense..

2) NeuCoin's paper appears to show that the number of possible iterations is SOO massive, that the search space is basically too large to.. err.. search.

Are both these ideas to fix short range attacks flawed ?

Because if not, well, POS chains branching off a POW chain would seem like a great idea..