Core Concepts

This page introduces the foundational ideas that underpin the Dualis Finance protocol. Each concept is covered in depth in its own dedicated section of the documentation; the goal here is to give you a complete mental model before you dive into the details.

Hybrid Lending

Traditional DeFi lending protocols operate in a single mode: over-collateralized. Borrowers must deposit more value than they borrow, and the protocol liquidates positions when collateral ratios fall below a threshold. This model works for anonymous, permissionless markets, but it is inherently capital-inefficient and excludes the credit-based lending workflows that institutional finance depends on.

Dualis operates a hybrid model that supports both over-collateralized and under-collateralized lending within the same protocol. Over-collateralized loans function similarly to Aave or Compound -- any verified participant can borrow against posted collateral with algorithmic interest rates and automated liquidation. Under-collateralized loans are available to participants who have been assessed by the credit scoring system and assigned a credit tier, allowing them to borrow at higher loan-to-value ratios commensurate with their creditworthiness.

This dual approach means that a single pool can serve conservative lenders seeking fully-secured yields and institutional borrowers who need capital-efficient access to liquidity -- without fragmenting liquidity across separate markets.

Asset-Agnostic Collateral

The Dualis collateral framework accepts three categories of assets, each with its own risk parameters and haircut model:

• Crypto assets -- standard digital tokens (e.g., CC, wBTC, wETH). These carry no additional haircut beyond the base LTV defined by the collateral parameters.
• Real-world assets (RWAs) -- tokenized securities such as government bonds, corporate debt, and equity. RWAs receive a 5% haircut to account for redemption latency and valuation uncertainty.
• TIFA assets -- tokenized institutional fixed-income assets originated through TIFA Finance. TIFAs receive a 20% haircut reflecting their illiquidity and structured payoff profiles.

By treating diverse asset types as first-class collateral with calibrated risk parameters, Dualis unlocks liquidity for tokenized assets that cannot be used as collateral anywhere else in DeFi.

Dual-Track Architecture

The protocol runs on a dual-track architecture that separates the computation layer from the settlement layer:

• Off-chain computation. Interest accrual, health factor calculations, oracle price aggregation, and credit scoring run on the Dualis backend (Fastify + PostgreSQL). This provides the throughput and flexibility needed for real-time risk monitoring across all positions.
• On-chain settlement. All state-changing operations -- deposits, withdrawals, borrows, repayments, liquidations, and collateral adjustments -- are executed as DAML contract exercises on the Canton Network. This provides deterministic finality, sub-transaction privacy, and an immutable audit trail.

The two tracks are synchronized through 32 Canton write operations that map backend actions to smart contract exercises. This architecture provides the performance of centralized computation with the trust guarantees of on-chain settlement.

Health Factor

The health factor is the primary risk metric for every borrowing position. It represents the ratio of a borrower's risk-adjusted collateral value to their total debt, determining how close a position is to liquidation.

Health Factor = Sum(Collateral_i * Price_i * LTV_i) / Sum(Debt_j * Price_j)

Where:
  Collateral_i  = Amount of collateral asset i
  Price_i       = Oracle price of asset i in USD
  LTV_i         = Loan-to-value ratio for asset i (adjusted by credit tier)
  Debt_j        = Outstanding debt in asset j (principal + accrued interest)
  Price_j       = Oracle price of debt asset j in USD

A health factor of 1.0 means the position is exactly at the liquidation threshold. Positions with a health factor below 1.0 are eligible for liquidation. The protocol targets a safe operating range above 1.5 and issues warnings as positions approach 1.2.

Reserve Factor

The reserve factor is the percentage of interest income that the protocol retains as reserves rather than distributing to suppliers. Reserves serve as a first-loss buffer against bad debt from liquidation shortfalls and fund protocol development and governance operations.

Each pool has an independently configurable reserve factor, typically ranging from 10% to 25% depending on the risk profile of the underlying asset. Higher-risk assets carry higher reserve factors to build larger safety buffers. The reserve factor is a governance-controlled parameter that can be adjusted through the proposal system.

Utilization Rate

The utilization rate measures how much of a pool's total supplied liquidity is currently being borrowed. It is the primary input to the interest rate model and reflects the supply-demand balance for each asset.

Utilization Rate = Total Borrows / (Total Supplied + Total Borrows - Total Reserves)

Where:
  Total Borrows   = Sum of all outstanding loan principal + accrued interest
  Total Supplied  = Sum of all supplier deposits
  Total Reserves  = Protocol-owned reserve balance

As utilization rises, borrow rates increase to incentivize repayments and attract new supply. The jump rate model defines a kink point (typically around 80% utilization) above which rates accelerate sharply to prevent full utilization and ensure suppliers can always withdraw.

Credit Tiers

Dualis uses a five-tier credit scoring system that determines borrowing terms for each participant. Credit scores are computed by the hybrid credit scoring engine, which combines on-chain borrowing history with off-chain institutional credit data from partner oracles.

Credit tiers affect three dimensions of a borrower's experience: the interest rate discount applied to base borrow rates, the maximum loan-to-value ratio allowed, and eligibility for under-collateralized lending. Tiers are reassessed periodically based on repayment history and updated off-chain credit data.

Canton Privacy

Dualis is built on the Canton Network specifically because of its sub-transaction privacy model. Unlike public blockchains where all data is visible to every participant, Canton ensures that each party in a transaction only sees the data that is relevant to their role.

In practice, this means:

• A supplier can see their own deposits, earned interest, and pool-level utilization statistics, but cannot see individual borrower positions or their collateral composition.
• A borrower can see their own loans, collateral, and health factor, but cannot see what other borrowers have borrowed or at what rates.
• The protocol operator has the visibility needed to compute interest accrual and monitor system health, but cannot unilaterally move user funds -- all state transitions require multi-party authorization via DAML contract exercises.
• Liquidators can see that a position is eligible for liquidation and the relevant collateral details, but only after the health factor crosses below the liquidation threshold.

This privacy model is enforced at the ledger level by DAML's authorization rules. It is not an application-layer feature that can be bypassed -- it is a fundamental property of the Canton Network's consensus and synchronization protocol. For institutional participants, this means their lending activity has the same confidentiality they expect from traditional prime brokerage relationships.