Comprendre comment les bridges fonctionnent

Types de bridges & le trilemme des blockchains

Cover Image for Types de bridges & le trilemme des blockchains



Vincent Bonnet



Types of bridges

Whether in an Ethereum/rollup-centric paradigm or in a rather L1-multichain one, bridges have taken a central role in the crypto ecosystem. They are indeed key for interoperability insofar as they allow different chains to communicate with each other and transfer value across different layers.

However, with the increasingly crucial role of brides also creates major challenges. How can bridges preserve blockchains advantages (trustlessness, censorship-resistance …) while fostering interoperability?

In order to do so, bridges require the following properties:

  • Trustlessness: Bridges need to ensure a security level equivalent to that of the connected blockchains
  • General usage: Bridges need to allow the transfer of value and data across blockchains
  • Extensibility: Bridges need to be able to connect different kinds of blockchain.
  • Have you already heard about the scalability trilemma? Here are the 3 pillars of the Interoperability trilemma. Bridges can only have two out of three features!


    (source: The Interoperability Trilemma)

    As trustlessness is a main concern for the crypto ecosystem, and the need for extensibility and generalizability differs depending to the use case, we are going to split the bridges into two categories: Trusted bridges and Trustless bridges.

    Before continuing with the different types of bridges, here is a quick explanation of bridges basics.

    Blockchain are not able to communicate with each other, and tokens only exist in one blockchain registry (ex: Ethers only exist on Ethereum blockchain). To use tokens on another blockchain and ensure communication between the two infrastuctures, we need to use bridges. Bridges work using smart contracts to create a representation of a token on another blockchain. To do so, they use a locking and minting mechanism to make the token illiquid on one chain, and to create the representation of this token on the other chain. The reverse operation consists in burning the token on a second blockchain and unlock the token on the first blockchain to make it liquid again.


    (source: What Are Blockchain Bridges)

    Trusted bridges

    Third party bridges

    Third party bridges are systems where transactions are verified off-chain. They are a set of external validators that operate communication between blockchains that are natively not-compatible. They can work with various mechanisms, being oracle based, MultiSig, MPC (Multi-party computation) systems, etc. As third-party bridges operate off-chain, they have less constraints. They can easily transfer any kind of data across different blockchains and be extensible to a lot of blockchains. This is the most common type of bridges to date.

    Exemple: Wormhole, Ronin, Multichain …

    Custodian bridges

    Custodian bridges are a kind of third party bridges that arefully trusted, as the cryptos are held by an intermediary such as a centralized crypto exchange. Custodian bridges’ role is solely to transfer value between different blockchains (no generalizability to other use cases).

    Example: Kraken, BitGo, Binance, …

    The biggest bridge is the custodian WBTC bridge operated by BitGo that give a representation of bitcoin on Ethereum. It accounts for roughly 40% of total bridge TVL at time of writing.

    Trustless bridges

    Natively verified (generalizable)

    Natively verified bridges are used to transfer any kind of data across chains with a similar technical stack (or domain. Ex: Ethereum domain, Bitcoin domain, Cosmos domain …). As they are not limited to value transfers, this kind of bridges are highly generalizable in terms of use cases. Natively verified and trustless bridges exist, for exemple, between Ethereum and its layer 2s (Starknet, Arbitrum, Optimism, etc.), or across Cosmos-chains. However, they are not extensible bridges because the way blockchains work in a specific domain is not easily transposable to another. They work indeed running a node of a blockchain with the VM (virtual machine) of another one to validate transactions.

    Exemple: Cosmos IBC, Rollup exits/entry mechanism (Arbitrum, StarkNet bridge …)

    Locally verified (extensible)

    Locally verified bridges are systems where only the parties involved in an exchange must verify the transaction, instead of relying on a broader verification system. These bridges can use different kinds of verification mechanisms. The most used are bridges based on atomic swap mechanisms. They work with peer-to-peer atomic swaps where the two parties cannot cheat (thanks to proof-verification mechanisms such as the one made famous by the Lightning Network on Bitcoin). These bridges rely on liquidity providers to benefit from a better liquidity and efficiency. Therefore, they are not generalizable because they imply economical counterparties. However, they are extensible because this mechanism is easily transposable to every blockchain as atomic swaps can be implemented with few requirements (possible on Bitcoin, too). Locally verified bridges can also work optimistically using layer 2 validators, to bypass the obstacle of the 7-days delay for withdrawals from Optimistic Rollups (ex. Arbitrum). More info about this here : What is Hop and how does it help Ethereum scale? (

    Exemple: Connext, Hop, Across V2, atomic swap mechanism …


    When we use different technologies, we always benefit from the security of the weakest element. As a matter of fact, even if we are bridging between two robust chains (ex: Bitcoin and Ethereum), the security depends on the bridges used, and in case of hacks funds are at risk. If funds are stolen from a bridge, wrapped tokens can lose their collateral and end up being backed by nothing. The choice of the type of bridge to use is important because it implies different security issues:

  • Technological risk: the risk of unexpected behavior in the code, or hacks
  • In the case of trusted bridges, there are additional risks:

  • Custodial risk: bridge operators can steal funds, or the custodial can be insolvent
  • Censorship risk: bridge operator can apply censorship to some users or to some transactions
  • However, using a trustless bridge does not necessarily imply more security. It is just subject to fewer types of risks. Bridges are indeed at an early stage, and they are very exposed to technological risk as shown by the numerous hacks targeting bridges.


    Bridge TVL Rankings - DefiLlama