Harmony’s Cross-Chain Future
Harmony was one of the first layer-1 blockchains to foresee and embrace the multi-/cross-chain future, where apps and assets will serve users on various blockchain platforms (multi-chain) and assets will flow across the blockchain boundaries (cross-chain).
Harmony’s Horizon bridge was one of the first cross-chain bridges that facilitated moving of assets from Ethereum and Binance Smart Chains to Harmony. The Horizon bridge was launched in October, 2020 and stood the test of time for nearly two years with total value locked (TVL) on Harmony crossed $750 million with over $1.5 billion transaction volume across 100k+ cross-chain transactions.
While the Horizon bridge was serving the Harmony blockchain in terms of enabling DeFi over cross-chain assets, Harmony pursued development of more secured and better trust minimized bridging technologies. While the Horizon bridge hack was definitely a huge setback for Harmony, we are strongly aligned with our belief of a better cross-chain future and our pursuit to contribute significantly towards the refinement of this technology. Below we share our knowledge of the cross-chain space, how we plan to adopt the various technologies, and also how we plan to contribute to its refinement.
A classification of various bridging technologies
At a very high level, there are two types of bridging solutions: 1) proof-based and 2) committee-based. The proof-based solutions entail cryptographic proving of the validity of any transaction of one chain in the other, whereas the committee-based solution relies on the social consensus of the bridge validators to attest for the transaction validity, which often involves no cryptographic proving.
Proof-based
In the proof-based solution, proving the transaction validity mainly concerns proving the block validity which contains that transaction because once the block is valid, the transactions contained within that block can be proved via merke proofs.
In the context of proof-of-stake blockchains, proving the block’s validity requires proving that the block has obtained consensus and has enough signatures attesting its validity. The block’s validity can be proved either via 1) validity proofs or 2) fraud proofs. The validity proving can be done naively via on-chain smart contracts or sophisticated technologies like zero knowledge proofs. The naive validity proving could be gas expensive, whereas it may be difficult/not-possible to apply zero knowledge proofs on certain cases. The fraud proofs on the other hand simply accept the block’s validity and allow a time window for any honest party to submit a fraud proof that invalidates it. Some of the notable examples of validity and fraud proof based solutions are listed below.
Types of validity proofs:
- Naive validity proofs: Replicates block validation on-chain. No additional optimizations required. Examples are: Cosmos IBC, Harmony’s Horizon 2.0
- Secure enclave with onchain proxy: Validation is performed inside a secure enclave and communicated to a on-chain proxy contract. E.g., Datachain’s Light Client Proxy (LCP)
- Zero Knowledge proofs: Succinct labs bridge between Gnosis chain & Ethereum, zkBridge
- Fraud proofs: Nomad bridge, Near’s Rainbow bridge
The naive validity proof based bridging solutions generally suffer from high gas cost in gas expensive environments like Ethereum. For instance, Harmony’s Horizon 2.0 technology incurs around 50 million gas to validate a block header due to the inherent cost involved in performing the BLS pairing check in a smart contract.
The cost will drastically decrease when the BLS operations are natively supported in Ethereum Virtual Machine (EVM)(refer to EIP-2537). Datachain recently proposed Light Client Proxy (LCP) which tries to off-chain the block validation to a secure enclave to reduce the cost of operation. However, this technology is yet to be fully developed. Similarly, the application of Zero Knowledge Proofs to block validation (under development) could not only optimize the cost, but improve security immensely.
In the absence of optimal validity-proof based bridging technology, fraud-proof based technologies have gained good traction, due to their single honest minority assumption, in which the security only requires a single honest entity that can watch over the validity of the relayed block headers and could submit fraud proof in the event of fraud.
Committee-based
The committee-based solutions involve social consensus between a set of validators without requiring cryptographic proving of transaction/block validity. The validator committee can be formed from the chain itself, where the cross-chain transactions are validated as part of the blockchain network. For instance, Cosmos Hub validators are responsible for validating Gravity bridge transactions, which connects Cosmos and Ethereum chains.
Types of committees:
- Built into protocol committee: Gravity bridge between Cosmos and Ethereum
- External proof-of-stake blockchain network or rollup chain: Axelar, Celer cBridge, Synapse chain (rollup chain)
- External non-blockchain networks with onchain smart contracts. Multisig: Harmony Horizon 1.0, Wormhole, Chainlink DON: LayerZero, Multi-Party Computation (MPC) network: Multichain
When compared to proof-based bridging solutions, the committee-based counterparts have their own pros and cons. The committee-based bridging solutions are more generalizable in terms of bridging between many chains, where proof-based solutions need additional development efforts to support more chains.
On the other hand, the trust minimization of committee-based solutions rely on the maturity of the bridge validator committee in terms of how decentralized the committee is in terms of participation cost (stake to become the bridge validator, machine cost involved), permissionless nature (can anyone with enough stake participate as bridge validator), collusion resistance (validator threshold or number of signatures and their total stake to take over the bridge, additional mechanisms to detect/prevent collusion), censorship resistance (if the bridge validators censor transactions, what mechanisms are triggered).
There are pros and cons to both proof-based and committee-based bridging solutions. The criteria that could be used to compare them can be:
- Trust minimization (in terms of participation, permissionless-ness, collusion and censorship resistance)
- Generalizability and customizability of the solution
Resuming Horizon
After the Horizon bridge hack, we thoroughly studied some of the proof-based and committee-based solutions. In terms of proof-based bridges, we considered Nomad, however the unfortunate incident of smart contract exploit of Nomad bridge was unfavorable.
In terms of committee-based bridges, we found LayerZero the most suitable solution in terms of non-disruptive and compatibility to Horizon bridge’s previous architecture. Meaning, it was easier to work around Horizon bridge’s contracts where we only had to swap out the multisig validation layer with LayerZero’s validation layer.
Note that, this is not our attempt to be judgemental about the various bridging solutions, but rather a subjective analysis keeping in mind our constraints in the context of Horizon. Beyond this article, there are several wonderful recent articles that layout the differences in various bridging solutions, like Jump Crypto’s article, ZK bridges, etc.
What next?
As previously highlighted, Harmony has put considerable effort into building a trust-minimized, cost/time-efficient bridging solution in the form of Horizon 2.0. However, the adoption of this technology is still dependent on the availability of certain cryptographic primitives like EIP 2537 in Ethereum mainnet. Such EIPs will eventually get deployed and in the meantime, Horizon 2.0 can explore various optimizations such as off-chain validation using secure enclaves as done in Datachain’s LCP and using Zero Knowledge Proofs as done in succinct labs project and zkBridge.
Harmony will also continue to adopt any superior bridging technology (in terms of trust-minimization and efficiency) that attains the required maturity over time to make them available for our users, as Harmony strongly believes in multi-/cross-chain future as well as multiple bridging technologies serving the multi-/cross-chain future.
The long term
Trust minimization will play a key role in shaping the future of cross-chain technology. Currently, neither the proof-based nor the committee-based bridging solutions are fully trust-minimized because of the lack of data availability. Meaning, both the proof-based and committee-based solutions require trusting the consensus of the connecting chains and it is not possible to easily validate the transactions and the state transitions (blockchain’s state) for an independent third-party.
This limitation can be resolved by making the transactions of the bridging chains available in data availability solutions such as Celestia or Polygon Avail. By making the data (transactions) available, any independent third-party can validate the cross-chain transactions and the corresponding state transitions (blockchain states) without needing to trust the set of validators that participated in the consensus of the bridging chains.
Note that, it is not just sufficient to make the data available for true trust minimization, but providing technologies to make the validation efficient. Meaning, any independent third-party can fetch the cross-chain transactions and reproduce the state transitions in a cost-/time efficient manner. A flavor of true trust minimization can be found in case of rollup bridges, where the rollup chain makes the data available in the main chain (like Ethereum) by posting its transactions as well as the corresponding state transitions (rollup state). However, the rollup chain is simply an extension of the main chain and hence the trust minimization in this context is within a single chain (naive).
The challenge of trust minimization across heterogeneous chains (with different consensus mechanisms) is still an open problem and the data availability solutions will play a major role in shaping the future of the trust-minimized cross-chain technology.
