Bridge

Basics

Projects, tools or companies who facilitate sending tokens between different blockchains. These bridges always come with trade offs. Most of the bridges are centralized.
Due to the different ledgers, different programming languages and blockchain technology being used within Layer 1 and Layer 2 blockchains, it is impossible for blockchains to seamlessly exchange tokens between each other. This created the issue of having siloed blockchain economies. To reach scale and to tap into each others liquidity, blockchain projects started building 'bridges' between each other. In practice, you cannot send ETH to Bitcoin, due to the completely different technology used (account model vs UTXO model), and even if it would be from EVM chain to EVM chain (like Ethereum to Polygon), its nodes would still not share the same ledger. A way to circumvent this issue, is by 'wrapping' tokens, creating a IOU of the original token, built in the code of the chain it gets bridged to. This means that the new wrapped token loses the native chain's attributes and instead has those of the new chain. More on the trade offs can be read lower in this page.
Coinbase wrote a good starter blog on Bridges and its risks (20-4-2022).

From a MakerDAO blog;

"Blockchain bridges enable interoperability between vastly different networks, such as Bitcoin and Ethereum, and between one parent blockchain and its child chain, called a sidechain, which either operates under different consensus rules or inherits its security from the parent blockchain (e.g., rollups built on Ethereum).1 This interoperability could include the transfer of tokens, data, and even smart contract instructions between independent platforms, allowing users to:

Deploy digital assets hosted on one blockchain to dapps on another
Conduct fast, low-cost transactions of tokens hosted on otherwise less scalable chains
Execute dapps across more than one platform."

Bridge Projects

Etherscan has a bridge dashboard.

A list of bridges can be found here.

Some of the well known bridge projects are:

Types of bridging

Atomic swaps

An Atomic swap is a smart contract technology that enables exchange of one cryptocurrency for another without using centralized intermediaries, such as exchanges. ... They first came into prominence in September 2017, when an atomic swap between Bitcoin and Litecoin was conducted.

From this presentation by REN (11-5-2020):

"Great for exchanging two assets, as long as you already know who you are going to exchange the assets with, you already know what assets you are working with and you already know the price and you are okay with either party being able to cancel the swap at any point. Therefor it is a pretty limited use case, and it has had limited adoption. They do not work for general applications and also not for actually every single blockchain either."

From this bridge blog (21-4-2022):

"This is arguably the most secure and trust-minimized method of transferring native assets to other chains, but this too has some drawbacks. Both chains must share the same hashing algorithm to ensure the blocks containing the transfer can be validated on both chains."

Liquidity Providers

From this bridge blog (21-4-2022):

"Pools are created on both chains and facilitate transfers for a fee. This system works fine when it’s adequately decentralized, efficiently re-balanced, and LPs have enough incentives to keep ample liquidity available on both sides of the bridge. Cross-chain swaps on this type of bridge are validated on both the source and destination chains. Assets are transferred from vaults typically managed by the LPs and validators.

It’s a big technical mess happening behind the scenes to make it easy for us to bounce around all these different chains, and whenever there’s money involved there’s potential for hacks and exploits. Vitalik warned us about some real challenges he sees with poor bridge architectures, potentially creating systemic risks. Colluding validators, compromised oracles, centralized relayers, and other dangers must be avoided when designing the way these tools function. In an interview discussing Celestia, Co-founder Mustafa Al-Bassam responds briefly to Vitalik’s assertion, saying he agrees with Vitalik for the most part. He has expressed these same concerns in the past. However, Vitalik’s argument is that every chain must use the same settlement layer to have shared security because you can’t do bridges safely. Mustafa believes this may still be possible with the right tools and that committee-based bridges can be an effective solution."

Lock and mint

From this bridge blog (21-4-2022):

"Lock assets on your source chain and mint synthetic tokens on your destination chain. This is the most common bridge type at the moment. ERC20 tokens are a great example. ERC20s can represent any asset on Ethereum. Bitcoin becomes $wBTC; even Ethereum itself can be wrapped to become $wETH.

Every network has a native token, and any network can issue its own version of another chain’s native token by “bridging” the asset. This is accomplished by locking the native asset on the source chain and providing proof on the destination chain that the assets are, in fact, locked. This allows the destination chain to confidently issue a wrapped version of the asset backed by the native assets locked in a vault on the source chain. The security of this system is dependent on the bridge and network of validators that validate transfers. Although it’s the most popular, concerns surrounding centralization, trusted custody, and collusion must be addressed to make this model more resilient. There’s also an inherent inefficiency that’s still being worked out. These bridges typically require a 2-step process to make synthetic assets available on the destination chain. Swap $ETH for a bridge token, then burn the bridge token to mint the synthetic asset. This presents an obvious opportunity for increased efficiency."

zk-Bridge

From a blog by Paul Veradittakit (16-2-2023):

"zk-bridges are altering how we think about asset transfers between blockchains. These bridges employ zk-SNARKs technology to allow for the efficient transmission of data or assets between multiple blockchain networks without the necessity of a centralized committee. A prover is used in these systems to rapidly demonstrate to a blockchain that a given state transition occurred on another blockchain. The blockchain keeps a summary of the most recent block on the other blockchain, and fresh blocks can be synchronized by creating and sending a zk-SNARK that verifies the other blockchain's tip has advanced from its previous state to its present state. Bridges built with ZK are also a step toward a future in which blockchains are more interoperable and can work together seamlessly.

Succinct Labs is developing a trustless interoperability solution that securely connects Ethereum and Gnosis Chain using on-chain light clients and zk-SNARK technology. The company is committed to a future of permissionless interoperability and has created a trust-minimized cross-chain bridge utilizing Proof of Consensus. The protocol verifies the source chain's consensus in the target chain's execution environment using on-chain light clients, which track block headers for each pair of communicating chains and rely on the economic security provided by the underlying L1 chain's consensus for security without making any additional trust assumptions."

Usage

From Etherscan (10-2021):

"Total balance held across all chains tracked in June amounted to $8 billion. Today, that number exceeds $18 billion. While Polygon’s total balance dropped from ~$7 billion to ~$5.5 billion, other bridges made huge progress in adding value. Only two non-Polygon bridges had balances surpassing $100 million in June; today three bridges hold more than $1 billion each and 15 have exceeded the $100 million mark.

As the main token deposited into Axie Infinity’s Ronin, AXS shot to the top of the list. Ether dropped to second place, partially explained by WETH appearing at #4 as several new bridges accepted it instead of Ether. Total AAVE amounts dropped, an interesting finding given that it was widely used in Polygon and is a major partner for the Avalanche Rush. Users may have been waiting for the launch to deposit into Avalanche. AXS is by far the most concentrated token per bridged chain out of the top tokens, with AXS locked in Ronin making up 56% of its circulating supply!

The combined $18 billion locked in Ethereum bridges exceed TVL for every single DeFi app across all chains, jumping from #8 four months ago."

Bridge Trade Offs

Bridges come in a variety of flavors with varying levels of trust associated with them.
Bridges are often able to freeze your coins.
Bridges often have their security handled by a company instead of by decentralized nodes.
Bridges do not bridge your 'real' token. But give you an IOU on the other chain. Look into wrapped tokens for more information or read further below.
Bridges are often touted as trustless, but in practicality are almost always permissioned and centralized. Users usually 'trust' the bridge creators with actually sending the user a representation token (wrapped token version) of the native token being bridged. This is due to the limitations of siloed blockchain technology. To have a bridge, the bridge creator deploys smart contracts on both chains, where the tokens get stored. Your native token never actually leaves its chain, it just gets stored into the smart contract. On the other chain, the bridge creator mints an IOU token of the locked native token. There is no direct communication that is decentralized between the two chains, since this would require the nodes of both chains to be in consensus, which they are not programmed to do. A bridge creator therefore often acts as middle man and 'checks' both chains for the user. This introduces trust. Some bridges are very clear in their centralization and are owned by a company or a project's Developer Labs. Other projects feign decentralization by putting the bridge's security in the hands of a group of chosen partners. So far it seems bridges can only be decentralized by putting a completely new consensus mechanism (with a new token) in between, like THORchain has done, thereby having a permissionless group of nodes handlining the security.
From this bridge blog (21-4-2022):

"Centralization — If only three people control a wallet or contract worth billions, the most likely outcome is those three people conspire to rob the bridge and ride off into the sunset.
Speed — It can take hours or days. If adequate liquidity isn’t present, you can find yourself waiting around a day or 2 for someone to bridge funds in the opposite direction or for destination side liquidity to be replenished.
Liquidity concerns — Bridges maintained by liquidity providers can run out of funds on one side of the bridge. This means users trying to bridge assets will have to wait for liquidity replenishment before seeing their assets show up on the other side of the bridge.
Smart contract risk — The assets that are locked on the source chain make the corresponding asset available on the destination chain. If the tokens locked in a vault on the source chain are stolen, the value of the wrapped assets on the destination chain goes to zero since there’s nothing to redeem them."