Introduction

Intent refers to expressing the desired outcome in blockchain transactions, rather than the specifics of the execution process. In traditional Ethereum transaction methods, users must specify every step of the transaction, including interactions with smart contracts, random number management, and paying gas fees. This method is often complex and inefficient. The introduction of intent aims to alleviate these burdens, allowing users to outsource the specific creation process of transactions to third parties while maintaining control over the transaction process.

In intent-centric transactions, users do not explicitly specify an execution path but provide a set of conditions that meet specific constraints. Users authorize third parties to choose the execution path on their behalf by signing and sharing their intents. Multiple intents can be included in a single transaction, allowing for matching overlapping intents to improve economic efficiency and gas utilization. Intents can also be applied to cross-domain transactions, enabling operations on different blockchains or systems and even allowing different payment methods for transaction fees.

Overview of Intent-Centric Architecture

Comparison with Traditional Transactions: In Ethereum’s traditional transaction methods, users must explicitly define how an action is executed and provide all necessary information for the Ethereum Virtual Machine (EVM) to perform state transitions. This can be complex and often leads to suboptimal user experiences and efficiency losses. In contrast, the intent-centric method focuses on the desired outcome rather than the specific process of achieving that result. Users express their desired outcome (the “intent”), while complex tasks are handled by third parties to find the best implementation method. This declarative approach improves user experience and efficiency.

Technical Functionality: Intents do not specify an exact computational path but allow any path that meets specific constraints. This flexibility allows users to maintain control over their transactions while outsourcing the creation of transactions. A single transaction can contain multiple intents, effectively improving economic efficiency.

Generally speaking, when submitting a transaction, users need to specify the exact computational path (execution steps). However, when submitting an intent, users specify a goal and some constraints, and the matching process determines the computational path to be taken. (Image source: Paradigm)

Intermediaries and Their Memory Pools

Intermediaries and their memory pools are crucial for the implementation of intents on the blockchain. Propagating intents in the Ethereum memory pool is the most obvious pathway, but the current design does not support the propagation of intents. Concerns about DoS attacks make the support for broadly applicable intents in the Ethereum memory pool a long-term challenge. Due to the open and permissionless nature of the Ethereum memory pool, additional barriers are created for the adoption of intents.

Without the support of the Ethereum memory pool, designers of intent systems face the decision of whether to propagate intents to a permissioned set or in a permissionless manner. A permissionless memory pool design might include decentralized APIs, allowing the propagation of intents among nodes in the system, but this approach faces challenges in terms of DoS resistance, propagation incentives, and Miner Extractable Value (MEV). On the other hand, a permissioned memory pool is more resistant to DoS attacks but may contradict the decentralized ethos of blockchain.

A hybrid solution might include permissioned propagation and permissionless execution, or vice versa, such as in order flow auctions. These designs require careful consideration of the quality of intent propagation and execution, ensuring that they do not become overly centralized or reliant on strong trust assumptions. Centralized applications involve not only new message formats for interacting with smart contracts but also alternative forms of memory pool propagation and counterparty discovery mechanisms. Designing an intent discovery and matching mechanism that is both incentive-compatible and decentralized is no easy task.

Source: paradigm

Intent flows from users to permissioned/permissionless and public/private intent pools, where it is converted into transactions by matchmakers and eventually enters the public memory pool or goes directly on-chain via MEV Boost-style auctions.

Applications and Examples

Widespread Application

In the blockchain domain, the concept of intent is simplifying the complexity of user interactions with the blockchain while allowing users to retain control over their assets and crypto identities. These intents relate to systems that have existed for years, including:

• Limit Orders: Users specify that 100 X tokens can be deducted from their account if they receive at least 200 Y tokens in return.
• CowSwap-style Auctions: Similar to limit orders, but rely on third parties to match multiple orders, maximizing execution quality.
• Gas Sponsorship: Allows paying Gas fees in USDC instead of ETH.
• Authorization: Allows interaction with certain accounts only in specific, pre-authorized ways. An intent is realized only when the final transaction adheres to the access control list specified in the intent.
• Transaction Batching: Allows batching intents to improve Gas efficiency.
• Aggregators: Execute actions only at the “best” price/yield. Such intents can be achieved by proving aggregation has been executed across multiple venues and adopting the best path.

The concept of intent is also garnering interest in cross-chain MEV, ERC4337-style account abstraction, and Seaport orders. For intent-centric applications, it’s crucial that at least one party understands the intent and is motivated and capable of executing it promptly. When designing intent-driven systems, considerations like the identity, method, and motivation of these executors are critical, as they determine the system’s effectiveness, trust assumptions, and broader impact.

Use Cases

Currently, intent-centric architectures in the blockchain domain are increasingly being applied, particularly showing their potential in decentralized applications (dApps). Here are some practical use cases:

COWSwap: Optimizing Transactions with Intent-Centric Methods

COWSwap, a renowned decentralized exchange, adopts the intent model to offer a more efficient trading experience. The platform introduces batch transactions and off-chain solvers to mitigate risks associated with Miner Extractable Value (MEV). Its core innovation is “COWs” (intents) - off-chain intents passed to solvers instead of individual transactions. These intents match with others that can balance the trade-off, reducing slippage. Additionally, COWSwap further develops this concept through multi-dimensional swaps, allowing multiple asset trades that don’t directly offset to be executed together.

Uniswap X: Finding Optimal Prices with Off-Chain Solvers

Uniswap X, another popular decentralized exchange, also harnesses the power of intent. The platform uses off-chain MEV searchers and solvers to find the best prices for swaps. By outsourcing the complexity of trade optimization to solvers, Uniswap X provides users with an improved trading experience. This approach not only simplifies user experience but also ensures traders obtain the most favorable prices in a highly dynamic and competitive market.

1Inch Fusion: Aggregating Liquidity with Intent-Centric Solutions

1Inch Fusion, a decentralized trade aggregator, also embraces an intent-centric design. It aggregates liquidity from various decentralized exchanges, optimizing the way users access the best prices and liquidity. Through intents, 1Inch Fusion ensures that users’ trade orders are effectively executed, even across multiple exchanges. This method simplifies the process of acquiring liquidity and securing the best trades.

The Role of AI in Enhancing Intents

Integrating Artificial Intelligence (AI) into the intent model represents a significant advancement in blockchain technology. AI can predict users’ intents more accurately by learning their historical preferences and behaviors, offering highly personalized experiences. AI also optimizes the matching of intents, ensuring users achieve the best outcomes. It analyzes vast amounts of data and market conditions to make intelligent choices. Moreover, AI enhances security by identifying and preventing potential malicious intents.

Benefits of Intent in Practical Use Cases

In practice, intent offers tangible benefits in multiple use cases. It simplifies complex transactions, saving users time and effort. By expressing preferences instead of specifying every detail, users can more effectively achieve their goals. Moreover, intent gives users the flexibility to define specific conditions and constraints, which is invaluable in scenarios like limit orders, smart orders, and cross-chain interactions.

These use cases demonstrate the practical application and potential of intent-centric architecture in modern blockchain technology. By intelligently matching and optimizing the trading process, these platforms enhance efficiency and user satisfaction. Furthermore, the integration of AI further strengthens the capabilities of the intent model, enhancing the intelligence and security of transactions. Overall, intent-centric architecture is becoming a significant trend in the decentralized application domain, offering users more flexible and efficient trading solutions. As this concept continues to evolve, we can expect to see more innovations and applications emerge in the blockchain technology and decentralized finance sectors.

Future Development Trends and Predictions

The future development trends and predictions of intent-centric architecture primarily focus on how to handle user interactions and meet their needs. This approach significantly differs from the current blockchain-centric model. Key focuses of the future intent-centric architecture include:

Defining and Focusing on User Outcomes

Intent inherently represents the desired end state of the user. This intent-focused approach marks a shift from the current blockchain infrastructure, often criticized for over-extracting value and not optimally serving end-users. Intent-centric architecture aims to empower users and decentralize power structures, thereby avoiding intermediary exploitation of end-users.

Generalized Intent Fulfillment Mechanisms

Unlike current architectures with mechanisms specific to applications, intent-centric systems offer a universal approach. This versatility allows them to address a wide range of user intents, offering flexibility and broader applicability. Key features of the intent-centric model include generalized intents, counterparty discovery, solutions, and settlements, all crucial for decentralized applications (dApps).

Advancements in dApp Development

The intent-centric architecture is expected to revolutionize dApp development. It introduces new features such as native and global scalability, information flow control, configurable ordering, and composite identities. These capabilities enable the realization of dApps that couldn’t be built with existing architectures.

Enhanced Transaction Models

In intent-centric systems, the focus shifts from specifying transactions to defining outcomes. This approach is more flexible and user-centered, allowing multiple potential solutions to fulfill user-defined end states, unlike the current more rigid transaction-based model in blockchain architecture.

Developing New Standards and Technologies

Entities like Essential are developing tools and standards to facilitate the adoption of the intent-centric model. This includes a universal Domain-Specific Language (DSL) for intents, standardizing the expression, combination, and resolution of intents. Additionally, intent-centric account abstraction standards for Ethereum and other EVM chains are being developed, integrating intent functionalities into the existing blockchain ecosystem.

Building Modular Intent Layers

Future developments include creating protocols with native support for intents. This intent-only architecture, different from the current Ethereum-compatible standards, aims to improve efficiency and reduce complexity by avoiding user-submitted transactions. Such protocols will also feature order flow aggregation and resistance to miner-extractable value (MEV), further optimizing user experience and reducing value exploitation.

Competitiveness Analysis

To compare the different approaches and effectiveness of these two architectures in handling user interactions and achieving user goals, a competitive analysis is conducted between Intent-Centric Architecture and Blockchain-Centric Architecture.

Intent-Centric Architecture aims to simplify the transaction process by focusing on user intentions, desires, and preferences, whereas Blockchain-Centric Architecture requires users to understand technical details, such as gas fees and smart contract coding.

Solvers in Intent-Centric Architecture optimize transactions to achieve efficient and cost-effective user intentions. Additionally, the features of Intent-Centric Architecture include the comprehensiveness and composability of intentions, enhanced cross-chain interactions, improved user control, and compatibility with Blockchain-Centric Architecture. It also addresses MEV challenges through innovative technologies, enhances the DeFi experience, strengthens privacy and security, and makes blockchain technology more accessible and user-friendly. Overall, Intent-Centric Architecture offers significant advantages in several key areas, especially in improving user experience, simplifying the transaction process, and enhancing cross-chain interactions.

Here is the competitiveness analysis table between Intent-Centric Architecture and Blockchain-Centric Architecture:

Risks and Challenges

In an intent-centric transaction architecture, several potential issues arise:

Centralization Risk: Widespread adoption of intents may lead to a shift of user activities to alternative mempools, which, if poorly managed, could result in centralization and rent-seeking intermediaries monopolizing the market.

Order Flow and Block Production Centralization: If the execution of intents is permissioned and the permission set is not carefully chosen, it may threaten the decentralization of Ethereum’s public mempool, leading to issues with block production centralization.

Challenges of Proposer-Builder Separation (PBS): Currently, the production of most Ethereum blocks relies on MEV-Boost under the PBS mechanism. If block builders gain exclusive access to transactions and intents (i.e., order flow), this could undermine the market structure that PBS relies on.

Censorship and Rent-Seeking Threats: Block builders controlling a large portion of Ethereum’s order flow could dominate the production of mainnet blocks, potentially leading to issues of centralized censorship. A single builder might shift value from Ethereum to themselves, posing threats of rent-seeking and censorship.

Middleware and Mempools: The propagation of intents poses significant design challenges. The current Ethereum mempool does not support the propagation of intents, leading to a dilemma in propagating intents between permissioned and permissionless systems. Permissionless mempools offer decentralized access but face challenges like DoS defense, propagation incentives, and MEV risks. Permissioned mempools perform better in DoS defense and handling MEV issues but contradict the spirit of blockchain decentralization, centralizing risks.

Opacity and Trust Issues: The intent-centric architecture requires users to relinquish some control over their on-chain assets, creating risks of opaque systems. This reliance on trusted intermediaries raises concerns about the fair distribution of power.

Further challenges in the intent-centric transaction architecture include the following:

Trust plays a key role in intent-centric architecture. Since many solutions rely on intermediaries, developing new intent-centric architectures faces a high entry barrier, potentially leading to reduced innovation and competition, thus impacting the assurance of execution quality.

In the worst-case scenario, if a single entity executes intents (like a monopolistic block builder), users lose their bargaining power. Moreover, the problem of centralized infrastructure in the market isn’t limited to the builder market. Even in non-block-building operations, high entry barriers may advantage intermediaries. Entities like Flashbots and CowSwap dominate most of the order flow auction market. Introducing new order flow auction designs requires establishing new trust relationships. These trust barriers to innovation and challenging the status quo may hinder the market’s healthy development.

Opacity is a major issue in intent-centric architecture.

When users transfer some control of their on-chain assets to intermediaries, especially in permissioned mempools, an opaque system may emerge. It becomes difficult to ascertain whether users’ expectations are met and threats in the ecosystem are detected.

This opacity is particularly critical for intent-centric applications that allow users to outsource significant decisions, like order routing. The negative impact of MEV on user transactions is often due to too much leeway given to executors (like slippage limits). Therefore, intent-centric applications that relinquish more autonomy need more careful design. In the worst case, using an intent-centric application means signing an intent that then gets translated into a transaction in an unclear manner. Even keen observers may struggle to monitor such an ecosystem, posing a threat to the health of Ethereum’s block production ecosystem.

Conclusion

To mitigate the risks in intent-centric architectures, it’s essential to consider the creation of an ideal system. Such a system should be permissionless, allowing anyone to match and execute intents while maintaining high execution quality. It should also be universal, eliminating the need for new memory pools when deploying new applications, and transparent, publicly reporting the process and data of intent execution and quality audit when privacy protections permit.

While teams like Flashbots and Anoma are diligently developing universal solutions that meet these requirements, the ideal system may not be ready in the short term. As a result, different solutions might serve different applications optimally. Although there may not be mechanisms like intent-specific crLists, small tools—such as allowing users to revert to traditional transactions when possible—could help improve worst-case scenarios. Hence, applications initiating intent pools should seek universality when permissionless and choose intermediaries cautiously when permission is required.

Overall, we urge designers of intent-centric applications to comprehensively consider the off-chain impacts of their applications, as these impacts might extend to the broader community, not just their user base. We also call for the wider community to remain vigilant about the development of the off-chain ecosystem surrounding Ethereum.