Bitcoin Sidechains
In this post I talk about what is a Bitcoin sidechain and why they have a key role in the cryptocurrency Ecosystem. Let’s start by briefly reviewing the history of sidechains.
The first attempt to build a sidechain dates back to 2012, much earlier than people think. The first technical paper that fully describes a Bitcoin sidechain was published in 2014 by Blockstream’s researchers. However the ideas on that paper were abandoned. In December 2015 RSK published its whitepaper, and began immediately working on the sidechain code based on a hybrid SPV-federated peg, with the potential to be extended to a drivechain, as presented in a BIP. In November 2016, RSK launched its first testnet network (connected to the Bitcoin testnet) and began testing their smart-contract applications with testnet bitcoins. Meanwhile, Blockstream started working on a federated peg, and published their design in January 2017. By 2017 RSK had performed several releases, and testnet resets. In January 2018, RSK finally launched its full-featured mainnet network. RSK native coin is BITCOIN and the RSK codebase does not mention any other token but BITCOIN. With the growth of the RSK ecosystem, many ERC-20 tokens were issued by the community, powering some of the most exciting DeFi applications. In September 2018 Blockstream launched the Liquid Network sidechain. Since then, no other Bitcoin sidechain has been launched to production, although there has been some advancements in drivechains, and a drivechain testnet has been launched. There has been other sidechain-like developments, such as statechains, but no testnet is yet available.
Blockchain Bridges
Two blockchains can exchange assets using a two-way-peg system, often called bridge. A bridge is a kind of communication protocol across ledgers. There are many kinds of bridges: Powpegs, SPV, hybrid-SPV, federated and collateralized. Each bridge has pros and cons. But not only sidechains build bridges. Many blockchains such as Ethereum, EOS or Polkadot use bridges to bring other coins into their platforms despite having their own native currencies.
A federated bridge consists of a set of functionaries that participate in a multi-signature custody of the pegged assets. The Liquid Bitcoin sidechain has a federated peg. There is an RSK/ETH federated peg as well.
In a Powpeg, the federation is replaced by a set of pegnatories, each running a special Hardware Security Module (HSM) that protects a private key of the multisig. Each HSM follows the blockchain PoW consensus to receive peg-out commands, therefore preventing pegnatories to have direct access to the pegged funds. RSK’s peg with Bitcoin is secured by a Powpeg.
In a collateralized bridge, the functionaries (called “vaults”) have full access to the private keys that control the bitcoins, but they need to deposit a collateral so if they steal the bitcoins in the peg, they are penalized by slashing their deposits. The amount of collateral must be greater than the pegged assets (overcollateralization) to prevent theft. If the collateral value is not enough because of rapid variations on the relative coin prices, an automatic liquidation process occurs. The consequence of this is that the collateral value must be much greater than the value of the whole economy enabled by the pegged asset. It’s surprisingly capital inefficient. One example of a collateralized bridge is PolkaBTC, which connects Bitcoin with Polkadot.
What it Means to be a Bitcoin Sidechain
Informally, a Bitcoin sidechain is an independent blockchain that can securely transfer bitcoins internally and from/to the Bitcoin network without supporting a money token different from Bitcoin.
Every sidechain needs a bridge to the parent chain. However not all of the previously described bridges are good for sidechains. In particular, token collateralized bridges, which require the issuance of another token, fit very poorly in a design of a Bitcoin sidechain. Due to the high collateral demand, any token used for collateralization of a bridge requires maximum stability to reduce the overcollateralization factor, and competes as a store of value. It must be a money token. Gresham’s Law stipulates that “good” money will be saved, rather than spent, pushing “bad” money to be used for payments. Therefore in the long run, no one will be willing to use “bad” money as collateral, because collateral can get stuck for long periods.
Why the Sidechain Definition is Important
The first definition of a sidechain (2012) can be found here. The sidechain is not defined by its function but as a blockchain with the following properties:
- Does not require separate currencies
- Does not need full or significant adoption by the bitcoin network
- Does not mirror the bitcoin blockchain block for block
- Uses and is dependent on bitcoin
During these years, blockchain researchers have attempted to come to a definition of sidechain and the debate is ongoing. In this period I sketched out several definitions of a sidechain trying to capture the original idea combined with what people now understands as sidechain, and I settled on the one mentioned in the previous section:
A Bitcoin sidechain is an independent blockchain that can securely transfer bitcoins internally and from/to the Bitcoin network without supporting a money token different from Bitcoin.
This definition has a technical part, but also refers to shared incentives. The definition does not use Blockstream’s interpretation in their 2014 paper, which is formally unclear.
When I say that a sidechain should not support a money token I don’t mean that a competitive money token cannot be created by the users on the sidechain platform, instead that a token should not be required for the network to function, nor incentivized by the sidechain community.
The most important thing about Bitcoin sidechains is that, no matter what definition we use, Bitcoiners will identify a sidechain when they see it. Therefore our responsibility is to preserve that meaning to avoid projects presenting themselves as Bitcoin sidechains only for marketing purposes. In short, the need for incentive alignment is key to the definition.
The fuzzy aspect of the definition is that the sidechain may support a token for governance or as an investment in future transaction fees, and nobody can prevent that token from becoming money. In that case the sidechain may cease to be so at a point in time. However, the presented definition of a sidechain is useful, generic and clear, and captures the need for aligned incentives without forcing the sidechain to use a particular consensus technology. But this definition is not widely accepted. For example, it’s not one that Vitalik agrees on.
Could the definition be improved? I don’t think so. But let’s explore some possible changes:
“Maximally decentralized”
By adding the adjective “maximally decentralized” to the definition we try to force sidechains to preserve the Bitcoin security model and prevent a centralized ledger to be called a sidechain. However, that would probably force us to drop the incentive alignment part, because all two-way-peg bridge protocols invented so far have security models that are strictly less decentralized than the blockchains they connect, especially when compared to Bitcoin’s proof-of-work consensus. A bridge that accept anonymous participants using another token as collateral may be more decentralized, but a a lot less incentive-aligned with Bitcoin. Apart from the Gresham’s argument, if we let the sidechain have a token for bridge collateralization, then some chains such as Polkadot, become Bitcoin sidechains, even that’s very far from the original sidechain meaning. So, to keep the original meaning, a sidechain should not support any other tradable token that tries to be money.
“Pays transaction fees in Bitcoin”
Adding this part it’s a slippery slope. The fact that transactions could be paid in bitcoin doesn’t mean users are incentivized to do so. Fees could also be paid in some other token at lower cost. Or the network could force fees to be paid in bitcoin but automatically convert them to another token to pay the miners. What if we require that they are paid and received in bitcoin? Then we must somehow make an exception for meta-transaction relayers (such as RIF Enveloping or GSN) that accept payments in stablecoins and buy the gas in bitcoin as a service to the users. I believe the use of fiat stablecoins does not interfere with Bitcoin money objective, as many users need to on-ramp with fiat currencies to be able to trade for Bitcoin for long term savings, and then trade back to stablecoins when they need stability over shorter terms. The objective of a Bitcoin sidechain cannot be to completely hide Bitcoin from the users, users need to be exposed to it. As a side-effect, forcing the definition to refer to transaction fees pushes some theoretical, but interesting, blockchains out of the consideration set. These include blockchains that don’t require the payment of a transaction fee and only rate limit user interactions, or have some other form of Sybil resistance mechanism such as per-transaction proof-of-work.
“No token pre-mine and no token subsidy in blocks”
In this case then a blockchain that has no single native coin, but allows several foreign cryptocurrencies to be simultaneously pegged to different native tokens and accepts all of them as payment for transaction fees would be a sidechain, even though the incentive alignment with Bitcoin could fade to none.
“Is secured by the Bitcoin mining-network”
There are a few consensus systems that rely on the bitcoin mining infrastructure: merge-mining, PoX and Veriblock. The most known, used and well studied consensus protocol is Nakamoto merge-mining. RSK uses Decor+ merge-mining for consensus and has achieved 80% of the Bitcoin hashrate. RSK researchers have developed an even more secure and advanced version of merge-mining called Inclusive Fork-aware merge-mining. However, new innovative consensus protocols are invented every year, with decreasing block latency. There is no reason to limit sidchains consensus to merge-mining, as Satoshi idea of maximum decentralization (“one-CPU-one-vote”) still lives in every bitcoiner. Therefore, while incentivizing the Bitcoin miners could be part of a sidechain’s incentive alignment with Bitcoin, it shouldn’t be its reason to exist.
Real Bitcoin Sidechains
The only existent Bitcoin Sidechains are RSK and Liquid Network. Both allow bitcoins to flow in, out and inside those platforms and none of them puts any other money token in a privileged position.
The relationship between Bitcoin and a sidechain is like when a cat and a dog are raised together. They are different, but they are part of the same human family.
Only that the dog is not a dog, but an indestructible honey badger.
Sidechains vs Remorachains
There are some blockchains that re-use some part of the Bitcoin network (generally to achieve consensus), yet they include a new token: we’ll call these blockchains remora chains. The Remora is a fish that attaches to another fish for its own benefit, but it neither helps nor harms its host. While these blockchains may have some technical merits, the existence of a separate, often competing non-Bitcoin money economy makes them non-sidechain and should alert bitcoiners.
The two existing Bitcoin remora chains are Veriblock and Stacks. Veriblock is a dead project but Stacks was relaunched in 2020 and supports a new language for smart-contracts. Both of them have their own native tokens. I don’t think the remora name is fully accurate because any new blockchain with a money token tries to capture a share of the store-of-value market and therefore works against Bitcoin. As long as it competes, it inflates the supply of store-of-value coins, while eroding the qualities of such. Therefore any token-collateralized remora chain is in fact to some extent harming Bitcoin (yet I won’t call them leech chains because they may contribute something else to the ecosystem).
To create a clear definition of a remora chain we must remove any system that uses Bitcoin but does not have its own blockchain, because that’s not “a blockchain”, but a chain of blocks. A Rollup contains a chain of blocks but it’s not “a blockchain”. There is an enormous difference in complexity and maintenance cost of a Bitcoin overlay protocol, which does not need a separate consensus algorithm to determine transaction ordering nor a separate peer-to-peer network, and an independent blockchain, which needs both. The main drivers for having an independent blockchain are two: faster blocks and bigger blocks (which equates to cheaper transactions), because onchain space is scarce and expensive. Common examples of overlay protocols are Counterparty, Omni, and other flavors of colored-coins. None of them I consider remora chains.
This is my definition of a remora chain:
A Bitcoin remorachain is an independent blockchain that can transfer bitcoins internally and from/to the Bitcoin network only with the support of a money token different from Bitcoin.
I don’t find remora chains especially attractive because generally there is a lack of common incentives between the remora and the host chains. While remora chains pay for some Bitcoin transactions, the average reward from those transactions is negligible compared to the average block reward. As a result, remora chains strongly rely on a host chain and a host community, but this community may turn their back on the remora chain at any time. It’s not a mutualistic symbiotic relationship. For example, Bitcoin remora chains usually use OP_RETURN payloads, which are restricted by a maximum size, which is enforced when transactions are independently transferred. This maximum size could be reduced at any time in the future with a simple software release, directly affecting the remora chain consensus.
A Creature Stranger than a Remorachain
Let’s imagine a blockchain which relies on many other host blockchains to secure consensus but the failure of any of them cannot prevent consensus. For example, a blockchain that can be simultaneously merge-mined with Bitcoin, Ethereum, Litecoin and zCash, and has bridges with all of them. This is not actually a remora chain, but a much stronger animal: maybe a wolpertinger.
Since currently there is no such chain, I’ll just leave it out.
Summary
The consolidation of Bitcoin as a store-of-value implies that sidechains are the natural extension of the Bitcoin finance stack, as Paul Sztorc has been repeating over and over since the beginning of time. A sidechain is a blockchain that is highly incentive-aligned with the Bitcoin community. RSK is the oldest and only Turing-complete smart-contract enabled Bitcoin sidechain, and already has a rich DeFi ecosystem. Bitcoiners define a Bitcoin sidechain as a blockchain that contributes to the Bitcoin ecosystem, and any definition of sidechain should preserve common belief. Bitcoin Sidechains are an inevitable part of the decentralized finance ecosystem. The increased growth of the RSK ecosystem suggests that the future of Bitcoin Sidechains is bright.
