This is the first in a series defining the Fabriq architecture, the goals, incentive structures, principles and complexities. Over the coming months we will dive deeper into our foundations but let us start at the beginning. In this blog post we will layout the immediate (but not only) problem we seek to address, the architecture at a high-level as it exists today in the Europa testnet, how we view the change in market dynamics, and where are we going from here. Here we aim for coverage, not completeness, and we’ll lay the groundwork for questions and subjects that will be explored further elsewhere.
The foundational problem of DeFi today is the intrinsic coupling of orderflow and execution. This arises in a diversity of ways. One such way appears in AMMs use of constant-function market making (CFMM) as the pricing mechanism. CFMM-based pricing was chosen in lieu of CLOB as it is much more computationally efficient but introduces an inherent dependency on the precise execution. CLOBs will generally match in batches reducing the dependency on ordering, while, given a single block, when the CFMM is called has a direct influence on the price you pay. It’s a little like a bad game of telephone given all the intermediaries (i.e., searchers) you have to go through to negotiation with the counterparty (i.e., the AMM).
RFQ (request for quote) systems have emerged as an attempt to address this issue by separating ordering from execution and providing a mechanism for counterparty discovery. However, these systems don't fully resolve routing concerns and still offer reliable signals for searchers to exploit, often binding counterparties to quotes that will be fulfilled at a loss. Generally, today's intent-centric systems fall short for similar reasons—users remain vulnerable to execution risks beyond their control. While solvers theoretically have greater capacity to manage execution, in practice, we've found this isn't the case.
These issues affect the overall transparency in the system. It’s true we can retroactively look at what has happened—however, constructing execution which will happen is out of reach for most participants in a blockchain. It’s very much not clear how to achieve preferred outcomes for most parties. In effect, there is a limitation on the agency of participants in the ecosystem to effectuate outcomes according to their preferences.
In short, Fabriq is a cross-chain intent orchestration and settlement protocol. We first need to outline some guiding design principles we strive to embody in our protocol design. In light of this, we’ll introduce the current architecture for our Europa testnet and its coming upgrade in Io, along with some remarks on areas this will be subject to change. Having the architecture in view, we want to posit effects and emergent properties. We’ll conclude with a view of the immediate future of Fabriq.
At this early stage of development and deployment, we want to clearly state our design principles and priorities. Our loyalty is not to a specific instantiation of a system or architecture, but a loyalty to principles that we believe we can contribute to meaningfully and push forward to the benefit of the ecosystem as a whole. Not withstanding, we still propose a style of architecture today which we believe fulfills these principles.
Incentive compatibility is a game theory concept where a game is structured so that all participants achieve their best outcomes by acting honestly. It comes in two forms: dominant strategy incentive compatibility (DSIC) and Bayesian-Nash incentive compatibility (BNIC). For our purposes, we're focusing on DSIC—we aim to design a system where everyone is incentivized to express honest preferences and act truthfully. Crucially, we want to constrain or entirely remove the blockbuilding ecosystem as a party in intent fulfillment.
As mentioned earlier in our brief overview of DeFi problems, current intent protocols have a vulnerability to incentive-incompatible behavior. Specifically, solvers face risks due to their reliance on AMM interactions. Without developing incentive-compatible liquidity sources, all such intent protocols will encounter similar challenges. The key to achieving incentive compatibility for all parties lies in constraining all system actors and minimizing the impact of execution ordering.
The Fabriq approach to incentive compatibility stems from the fundamental separation of execution and state construction from order flow and counterparty discovery inherent in intents. To extend incentive compatibility further, we must expand the use of intents in applications. If liquidity itself is exposed in an intent-centric manner—unlike today's AMMs—then solvers (and LPs) can move closer to eliminating the externalities of blockbuilding. Given the prevalence of complex adversarial transaction re-ordering schemes, we believe intent-centricity is the future for many applications, extending beyond just liquidity provision.
We believe universal fungibility is an implicit ideal of the blockchain ecosystem: the ability to exchange anything, with anyone, anywhere, at any time. We want every conceivable exchange of value to be possible. While this concept isn't particularly novel—it's implicit in why many of us are building in this space—it's crucial to understand its importance.
To define universal fungibility more formally: Consider a set of assets, S. Universal fungibility requires that for any two assets s_1 and s_2 in S, exchanging one for the other and then back again should yield approximately the same value as the input asset (excluding the cost of the exchange itself). Alternatively, and perhaps more practically: given a reference asset r in S, universal fungibility demands that for all s_1 and s_2, swapping s_1 for r, then s_1 for s_2, and finally back to r should result in approximately the same quantity of r, our reference asset.
Clearly, universal fungibility is an ideal that can't be fully achieved. For one, there's little demand to exchange assets without functional markets (like pet rocks) for those with actual value. Moreover, factors such as fees, token support, and network differences naturally impact the achievability of universal fungibility.
Let's illustrate this with a simplified, albeit convoluted example. Imagine a trader wanting to move ETH-ARB → OP-USDC → POLYGON-USDT → ETH-ARB—an admittedly foolish round trip that no one would begrudge incurring hefty fees. Ideally, we'd expect this idempotent series of exchanges to return us close to our starting point. In reality, a significant portion would likely be consumed by fees. The ultimate goal in pursuing universal fungibility is to address these smaller issues and align incentives to enable smoother exchanges: improving liquidity pricing, enhancing chain security, and resolving finalization issues, among others.
Fabriq’s ability to achieve universal fungibility also largely isn’t something novel. It’s a commitment to working to align the incentives of all the participants of the network. We want to ensure solvers have access to valuable opportunities. We want to ensure intent-centric applications have access to a diversity of networks, reliable execution and deep liquidity. For the users, we want to instill confidence that their preferences will be faithfully executed. This only comes from dedicated and careful cultivation.