Intention System: A Solution to the Complexity of DeFi?

Before the collapse of Luna, I managed a stablecoin yield strategy for a friend who wanted to understand the high interest rates available in January 2020. My friend was not familiar with cryptocurrencies and had never conducted on-chain operations before. Our collaboration was simple: he would deposit funds into a hardware wallet, and we would meet via Zoom once or twice a week, where I would gradually guide him through the process.

We distribute funds across multiple DeFi protocols on various chains. In a session lasting 2-4 hours, we conduct dozens of transactions, including approvals, transfers, swaps, deposits, claims, and withdrawals. Funds are transferred into customized liquidity pools, voting locks, etc., to maximize incentives. We utilize various cross-chain bridges, mainstream decentralized exchanges, yield aggregators, etc., to capture any yield in the cryptocurrency space.

This process is quite complex. I need to explain each step in detail, while my friend needs to follow the instructions and familiarize themselves with the interfaces of various DeFi tools. The meeting was filled with instructions like "click here", "go there", and "swap this". For example, to exchange USDC for FRAX/DAI LP on Polygon, it requires 12 transactions:

1. Exchange USDC for DAI on DEX ( 2 transactions )
2. Cross-chain USDC and DAI to Polygon(4 transactions)
3. Merge USDC and DAI in a transaction on Polygon DEX ( )
4. Deposit LP into the vault to receive rewards (2 transactions )

This simple flow of funds requires 12 transactions! We have to manually search for, create, and execute each transaction, a process that is lengthy and cumbersome, especially for large-scale portfolios.

From a higher level, all the processes we execute have clear expected outcomes. We have assets and want to use them to accomplish specific tasks. Just like the example above, "We have USDC on Ethereum and want to provide liquidity in the form of FRAX/DAI on Polygon, and then deposit it into the staking vault." This is the "content" of the operation, while the 12 transactions represent the "how" of the operation. A series of clear and logical steps are needed from start to finish.

If there are powerful algorithms to handle transaction routing, this process can be simplified to 1-2 steps. We just need to specify the desired outcome, and the algorithm can return the optimal path, even directly processing the transaction. This path mapping structure is called "intents" and is part of the rapidly developing middleware future of Ethereum.

Although there is no consensus in the industry on the definition of "intention", there are some general concepts. Paradigm defines it as: "An intention is a signature of a set of declarative constraints that allows users to outsource the creation of transactions to third parties while maintaining complete control over the transaction." David Ma from Near believes that: "A transaction is imperative, while an intention is declarative. A transaction is a well-defined message that specifies how to operate the EVM to produce state changes, while an intention specifies the desired state change without focusing on the implementation process."

Both definitions emphasize that intent is "declarative", seeking external help through data sharing between the user and the "solver". The user declares the desired outcome, and the solver provides a means to achieve it. Unlike transactions with specific parameters, intent requires a third party to map it. There are also constraints that limit the possible set of paths. This helps to narrow down the possibilities to a smaller, filterable set for the user to choose from.

In the case of my friend's example, the intent system allows us to broadcast the final goal to a group of solvers, which calculate the optimal path. We then choose the route with the best price and execute the trade. All intermediate steps are handled by the solvers, and the user only needs to confirm 1-2 transactions.

The basic architecture based on "intention" already exists in the EVM. When using any decentralized exchange, it will find the best trade execution route. In the Curve interface, after selecting the assets to buy and sell, the UI will automatically find the best liquidity pool for routing. Since there is no direct trading pair for USDT/frxETH, the order will go through multiple pools to achieve the best execution path: USDT > sUSD > sETH > ETH > frxETH, all completed in one transaction. It will also roughly provide price impact and how to limit slippage. After selecting parameters, the UI can also help construct the raw EVM data for broadcasting.

This Curve trading intention is just a basic example. The UI is a useful tool for building swap trades, and the logic is as follows:

FRAX Trading: Exchange 100,000 FRAX for at least 999,000 USDC using the 5bps FRAX/USDC Curve LP, valid until block X.

In contrast, the intent system will share the expected outcome ( to acquire the maximum USDC ) and the constraint ( to only sell 100,000 FRAX ). It is up to the solver to determine the optimal exchange rate.

If you have used 1inch or DeFiLlama, you will see the intent system used to build exchange trades. When using Llamaswap, you still need to provide all execution parameters and then obtain a set of potential trade relayers. Most of these exchanges are executed on Curve except for (CowSwap ), but the fees and gas costs vary depending on the relayer. Ultimately, it is up to the user to choose the best price/cost combination.

In addition to trading aggregators, there are several other types of "intent" on Ethereum:

1. Limit Order: Allows asset withdrawal from the account when conditions are met.
2. CowSwap-style auctions: Third-party order execution based on non-DEX liquidity
3. Gas Sponsorship: Allows the use of tokens such as FRAX to execute third-party transactions, applicable to account abstraction wallets.
4. Delegation: Whitelisting belongs to this situation, check the database before executing the transaction.
5. Transaction Batch Processing: Allows for batching of gas efficiency intentions.
6. Cross-chain exchange: See Socket

Although the types of orders are becoming increasingly diverse, the simplest way to describe the intent may be "limit order," which is just a new marketing term. A limit order refers to the desire to purchase a specific quantity of an asset at a specific price, and it will only be executed once the other party accepts the order.

Similar to limit orders, the intention consists of two parts. The first part is the desired final state of the user, and the second part is the transaction initiated by the solver. By combining the two, all the necessary elements for executing the transaction can be obtained.

Sell MEV

The construction method based on intent architecture carries almost no risk. First, solvers have no incentive to propagate MEV intents that they can profit from. "In many cases, extracting MEV requires executing user orders on-chain. In such cases, the execution of user orders exposes the blockchain state, which extractors can exploit for profit. Backrunning and sandwich trading are common examples."

The core feature of intent is data exposure. Signing an intent message indicates your willingness to extract MEV for convenience. Since intents cannot be directly broadcast to the Ethereum mempool ( transaction execution's queuing area ), they are filled in a private off-chain Interpool. These Interpools can be permissioned, permissionless, or a mix of both.

Permissionless Interpool adopts a decentralized API, where nodes in the system can freely share intentions and grant executors unrestricted access. For example, the 0x protocol relayer and the proposed shared ERC4337 memory pool. The open memory pool is susceptible to DDOS attacks and cannot ensure the prevention of malicious execution intention propagation.

In contrast, the permissioned memory pool uses trusted APIs to resist DDoS attacks without the need for intent propagation. By relying on trusted intermediaries, as long as trust is maintained, the quality of execution can be guaranteed. Such intermediaries are usually reputable, which motivates them to ensure top-notch execution. However, they still operate under strong trust assumptions, which contradicts the core spirit of open blockchains.

Hybrid solutions bridge the gap between permissioned and permissionless systems. They may combine permissioned propagation with permissionless execution, or vice versa. Order flow auctions like the CoW Protocol use a trusted party ( protocol for off-chain order matching ) to operate the auction, but participation is permissionless.

The currently most popular Interpool is centralized and permissioned, with no incentives to share information with competitors. The risk is that one party absorbs most of the intent-based transactions and uses its monopoly position to impose fees and engage in other rent-seeking behaviors, leaving users with negotiation power disappearing into the exploitative middlemen.

middleware risk

When considering intent as a limit order, we can compare the order flow of certain trading platforms with the payment (PFOF).

These platforms offer users "free" trading, based on the premise that users can sell order flow instead of sending it to traditional exchanges. Market makers are companies that buy and sell securities in large volumes, and they provide this payment because they can profit from the bid-ask spread of the orders. Critics widely condemn this practice due to conflicts of interest. While brokerage firms are obligated to provide best execution for client orders, the monetary incentives of PFOF are said to influence their decisions on where to send the orders.

The intention is a form of PFOF arbitrage, which we call MEV. The arbitrage opportunities created by the long-term open orders ( and partial orders ) may be more valuable than transactions manually added to the Ethereum mempool, because solvers can determine the route instead of competing with sandwich trades for pre- or post-MEV in a given block.

Unchecked and opaque solvers are likely to provide the worst routes, as their profit margins are inversely proportional to good execution. Users still need to choose solvers, and they can leverage this negotiating power to force solvers to compete for order flow. The solver that provides the highest return to users under constraints wins the auction.

Some protocols adopt this design, using batch auctions to find the best settlement price for traders. In these protocols, orders are not executed immediately but are collected and settled in batches. The system does not use a central operator, but instead utilizes public competition among solvers to match orders. After the batch concludes, these solvers submit solutions for settling the orders.

Batch auctions allow transactions within a batch to have the same price, thus eliminating the need for miners to rearrange transactions. There are no pre-run or post-run. These protocols use order flow auctions to ensure that traders achieve the best price execution. However, there is some MEV in the orders, as market makers must be able to perform arbitrage trading in another venue to remain profitable.

Future

Currently, some protocols are developing intent-based infrastructure to allow hybrid systems. Certain projects are building private memory pools and block building networks to direct traffic to L2 and Ethereum. There are also projects attempting to build the next generation of fully permissionless infrastructure, and several other companies have joined in.

Although there is no consensus on who the winner of the intentions is, this is an increasing part of the emerging middleware layer revolution happening in the crypto space today, which is necessary for convenience. The current UI is not user-friendly enough for widespread adoption. The current intentions are typically used for currency exchanges and order batching, but the goal is to make them applicable to completely general data and arbitrary data.

This opens the door to new possibilities, as all wallets may default to account abstraction. A powerful intent layer can unlock new use cases for products and simplify the applications built on it.

![Why Intents might be the answer to the complexity issues of Decentralized Finance?](