Introduction to interoperability The interoperability of blockchains has been a central topic for a long time and continues to grow in importance in an increasingly multichain world.
Within a blockchain, it is possible to carry out actions in a "trustless" way, i.e. we are certain (or almost certain) that they will happen because they depend on the code that governs a network whose architecture forces the actors operating it to coordinate. Going against this code and compromising the network would be almost impossible and above all unprofitable.
However, a problem arises when we try to get two blockchains to communicate with each other, since a priori they cannot trust each other.
Several solutions attempt to address this problem.
Trusted bridges: Wrapped Bitcoin and Circle's CCTP A group of players can be trusted to transfer cryptos from one chain to another. This is useful, particularly when you want to make a blockchain like Bitcoin (which lacks programmability) communicate with another.
To address this, a group of centralised players (mainly Bitgo, Kyber Network and Ren) have come together around a DAO to issue WBTC (Wrapped BTC, i.e. bitcoins 'packaged' on another blockchain).
This takes the form of a synthetic asset, present in particular on the Ethereum, Base, Osmosis and Tron networks. The rule is that for each WBTC created on one of these chains, there is a parallel bitcoin held by custodians on the Bitcoin blockchain, so that it is always possible to claim 1 BTC in exchange for 1 WBTC.
Currently, there are 153,200 WBTC in circulation (around $10 billion).
Although there are audits of accounts and efforts at transparency, this system is far from perfect, as it is not automated and requires trust in a small group of players.
However, centralised bridges are not always as problematic, particularly when they transfer tokens that are also centralised.
Circle, the company that issues the stablecoin USDC, has implemented the Cross Chain Transfer Protocol (CCTP), which allows USDC to be transferred between supported chains (Arbitrum, Avalanche, Base, Ethereum, Noble, Optimism, Polygon PoS and Solana).
CCTP is based on a "Burn & Mint" system made possible by Circle being the official issuer of USDC. Thus, each USDC transferred through CCTP is always considered a canonical (non-synthetic) asset. Using CCTP therefore does not introduce any additional trust assumption in relation to simply holding the USDC.
General message transmission infrastructures : Wormhole, Layer Zero, Hyperlane, etc. General message transmission infrastructures allow the chains they integrate to exchange any type of message. On these infrastructures, it is possible to build all sorts of applications such as DEXs, cross-chain lending protocols and above all bridges.
Among these projects, we can name Axelar, Wormhole, LayerZero (read our project analysis) and the recent Hyperlane.
The interest of Hyperlane is to allow any blockchain to connect to its network without permission. This means that blockchain projects no longer need to spend time negotiating with interoperability infrastructures in order to connect to them, which is handy in a world where the number of blockchains is growing every day.
Each infrastructure brings its own standard for creating tokens that can be easily transferred from one chain to another with "Burn & Mint" systems.
These infrastructures involve additional assumptions of trust since very restricted committees of operators must be trusted.
However, restaking (read our survey) helps to reduce this problem by strengthening the economic security of these infrastructures. For example, Renzo (read our project nalysis) relies on the security provided by EigenLayer (read our project analysis) to ensure ezETH cross-chain transfers, as Hyperlane is an EigenLayer AVS.
Oracles: Chainlink and its CCIP Oracles are entities that allow a blockchain to communicate with the outside world, including other blockchains. In this way, oracles can compete with infrastructures in the previous category.
This is in particular the ambition of Chainlink, which since mid-2023 has been developing the Chainlink Cross-Chain Interoperability Protocol (CCIP), capable of transferring all types of messages and cryptos between blockchains.
The limitation of this approach is the small number of integrated networks compared with the previous category (only EVM chains, including Ethereum, Optimism, Arbitrum, Polygon, Base, etc.), as well as the cost of transfers, which is higher.
By way of comparison, Hyperlane manages Cosmos and Starknet in addition to EVM. Layer Zero also supports Aptos and Solana. And Wormhole can also connect to Solana and Sui.
Chainlink's solution is struggling to take off, as only two minor projects appear to have adopted it so far.
The Cosmos IBC The Inter-Blockchain Communication Protocol (IBC) is a set of interoperability standards built by default into chains built with the Cosmos SDK, the common development kit for the Cosmos universe, but it can also be extended to other chains.
IBC enables any type of communication between blockchains, whether it be crypto transfers or message exchanges.
To put it simply, the IBC relies on a set of "light clients" (nodes that check the state of a blockchain without downloading it in full) that ensure that messages transported by "relayers" (off-chain entities) are correct.
If the relayers attempt to transmit erroneous messages, these will be refused by the light clients. This design choice underlines the trade-off of IBC, which prioritises security over business continuity.
IBC is one of the most secure means of interoperability and has proved its worth so far, but it is not immune to vulnerabilities for all that.
The limitations of IBC include the number of light clients a blockchain must integrate to communicate with other connected blockchains, as well as the cost of verifying messages exchanged, which can be very high, particularly on Ethereum. The Union protocol aims to solve these issues by coupling IBC with zero-knowledge proofs.
Layer 2 canonical bridges Layer 2s are all connected to their layer 1 through a canonical bridge, which is a kind of official bridge.
A canonical bridge has the same level of security as the project that deploys it. Using cryptos transferred through these bridges therefore adds no additional trust assumptions on top of those inherent in using the layer 2 in question.
One of the limitations of canonical bridges is their slowness, intended to ensure maximum security. They have to wait until a transaction is irreversible before transferring. This often takes several days.
"Intents": Across and Everclear "Intent"-based bridges allow users to issue an intent to transfer value. A group of off-chain entities called "relayers" or "solvers" then compete to fulfil this intention, offering the best possible price and the shortest possible timeframe.
These relayers are often sophisticated players, such as market makers, who possess liquidity on several chains and accept the risk that a transaction may be cancelled. This transfer system places the risks on the relayers, who are reimbursed only after users have been served. Across is one of the most significant bridges using intents (read our analysis) .
Several advances will continue to improve the efficiency of relayers and therefore the bridges they serve.
The first is the introduction of standards, enabling relayers from different projects to communicate with each other, such as the ERC-7683 developed by Across and Uniswap.
The second is the development of a clearing house for intent-based protocols. Developed by Everclear (formerly Connext), it will greatly improve liquidity management for relayers.
Agglayer: Polygon's vision Agglayer is the interoperability solution developed by Polygon, to which any channel (layer 1 and layer 2) can connect (read our analysis of Polygon and Agglayer).
The Agglayer consists of two components:
A unified bridge on Ethereum allowing all bridged assets between chains connected to the Agglayer to be fungible with each other. A verification and aggregation system for zero-knowledge proofs (ZKPs). ZKPs have the property of being succinct, meaning that they can be used to verify an unlimited amount of data very quickly. ZKPs therefore allow two chains to communicate without introducing a trust assumption thanks to an exchange of mathematical proofs of their states. ZkSync uses a similar architecture to develop its "Elastic Chain".
The current limitations of ZKPs are their cost and generation time. Enormous progress has been made in recent years, and these improvements should accelerate with more and more projects in this sector and the development of specialist hardware.
The Big Whale's opinion Some projects try to transfer all types of messages (Wormhole, Layer Zero and Hyperlane), the problem being that their architecture is relatively centralised at the moment, the integration of restaking partly addresses this problem, and their operation is often quite slow.
Their main strength is their ability to create standards for transferable tokens without friction between chains, the problem being that each brings its own standards.
On the other hand the protocols that rely on intents (Across and Everclear) focus solely on value transfers and seem best placed to transfer all cryptos that are not created with the previous standards. The advantage is that they are able to accept similar standards and improve their synergies.
In the medium to long term, the reduction in ZKP generation costs and time should enable chains to communicate with each other quickly and trustlessly and therefore replace current infrastructures in most cases. Thus, Polygon and Union's approaches seem the most promising.
