Skip to content

Latest commit

 

History

History
204 lines (150 loc) · 20.8 KB

README.md

File metadata and controls

204 lines (150 loc) · 20.8 KB

Group 3

  • Capocasale Vittorio
  • Castellano Dario
  • Renieri Margherita
  • Toffalini Alessandro

Algorand NFT Loan System

Environment setup

Create the development environment:

  1. Connect to a network
    • Using Sandbox
      • Make sure the docker daemon is running and the docker-compose is correctly installed
      • If you do not have the sandbox cloned, follow the Sandobox installation guidelines otherwise go to the next step (launch it).
      • Launch sandbox with ./sendbox up dev
    • Using Testnet
      • Create an account on PureStake and get an API key
      • Export in the environment the following variables:
        • ALGOD_TOKEN with your Purestake API key
        • ALGOD_ADDRESS with the address of the Purestake testnet algod
  2. Clone the repo
  3. Check your python version is >= 3.10. Verify it with the python --version
  4. Create the Virtual Environment
    • pip3 install virtualenv
    • virtualenv venv
    • source venv/bin/activate
    • pip3 install pyteal py-algorand-sdk
  5. Test the environment with https://developer.algorand.org/docs/sdks/python/
  6. Install all the required packages
    • pip install -r requirements.txt

Goal of the project

The project aims to provide liquidity in terms of cryptocurrency or specific assets that could be addressed with a loan mechanism of owned NFT.

Use case

The developed system can be seen as a kind of virtual pawnshop. The use case involves a user who has an asset, in this case an NFT, that he does not want to deprive himself of but at the same time has a need for cryptocurrencies. Through our smart contract, the owner of the NFT has the opportunity to receive a loan, allocated through an auction system, from a pool of potential investors using his NFT as collateral. At the end of the predetermined period of time he can then repay the debt with due interest and have his NFT back.

Smart contract specifications

In the NFT loan developed, there are three different roles:

  • The Contract Owner (CO) that creates and inizialize the contract
  • The Lender (L) that provides assets to the smart contract
  • The Borrower (B) that provides her NFT as collaterla to borrow cryptocurrency.

Once the contract owner created the smart contract, it is able to accept an NFT from an account - borrower - for a period of time and stores info on the global state. During this period other accounts - lenders - are able to bid in ALGO for that NFT. At any time during this initial period, the borrower can interact with the contract to get back the NFT (in this case the bidded amount returns to the lender) or to accept the loan.

In the second case, there is a predefined period of time in which the borrower must repay the loan to the lender, otherwise, after the deadline, the lender can obtain the NFT by interacting with the contract.

The Stateful Smart Contract stores the following information:

- nft_id (a handle to retrieve the correct NFT)
- borrower_address (the address of the borrower)
- lender_address (current highest bidder)
- highest_bid (highest bid)
- auction_base (auction staring amount)
- auction_period (auction deadline block number. The client submits a duration, the contract computes the deadline)
- payback_deadline (loan deadline. It initially stores the duration. It is calculated after the accept_offer is invoked)
- last_interest_update_block (starting block to compute the interest. It stores the block corresponding to the last successful invocation of pay_back)
- debt_left (current debt. debt_left=debt_left*((1+interset_rate)^(current_block - last_interest_update_block)))
- state (current contract state)

The smart contract supports the following operations:

  • provide_access_to_nft(NFT, payment)

provide_access_to_nft is sent by B to the smart contract to trigger the NFT optin. B must provide an atomic payment transaction to address the minimum balance requirements (0.1 Algo).

  • set_offer (NFT, auction_base, auction_period, payback_deadline)

B sends the NFT to the smart contract through an atomic asset transfer transaction. B establishes a minimum loan threshold (auction_base), the number of blocks of the auction validity period (auction_period), and the loan payback deadline (payback_deadline), which is the number of blocks from when accept_bid is invoked. The smart contract stores B's address (borrower_address) for future ownership transfers.

  • place_bid (n_Algos)

place_bid is invokable only during the auction validity period. L sends some Algos (n_Algos) to the smart contract through an atomic payment transaction. The smart contract stores the lender's address (lender_address). The amount of Algos must be greater than the loan threshold and the current highest bid. The smart contract refunds the previous highest bid and replaces it with the new bid.

  • accept_bid ()

accept_bid is invokable only by B. The smart contract keeps a small percentage of the n_Algos and forwards the remaining part to B. The percentage represents a service fee that CO can collect by invoking the pay_me function. The smart contract still owns the NFT. cancel_offer and timeout can no longer be invoked.

  • timeout ()

timeout is invokable only after the auction ends and if accept_bid is not invoked.
Anyone can invoke timeout to return the managed assets (NFT, n_Algos) to their original owners.

  • cancel_offer ()

Only B can invoke cancel_offer, and only if accept_bid is not invoked. The smart contract returns the managed assets (NFT, n_Algos) to their original owners.

  • pay_back (m_Algos)

B gives some Algos (m_Algos) to the smart contract through an atomic payment transaction. pay_back updates the current debt by summing the accumulated interest. m_Algos must repay at least the accumulated interest.

If m_Algos exceeds B's current debt, the smart contract returns the exceeding Algos to B. The smart contract subtracts m_Algos from B's debt. The smart contract forwards the the m_Algos to L.

If B's debt goes to 0, the smart contract gives the NFT back to B. pay_back can be invoked after the loan payback deadline expires but not after loan_expired is invoked.

  • loan_expired ()

loan_expired can be invoked only by L after the loan payback deadline expires. L receives the NFT from the smart contract. pay_back cannot be invoked anymore.

  • pay_me ()

pay_me can only be invoked by CO. pay_me sends the currently collected fees to CO's address.

State of the art

Public attention towards NFTs has rapidly risen in 2021, when NFT market has experienced record sales. In just one year, it has been shown that NFT market increased from total daily sales of about USD 183,121 in 2020 to an average of USD 38 million at the end of 2021. According to DappRadar, the third quarter of 2022 saw USD 3.4 billion in NFT sales. From the current trend, it is evident that the digital asset market will be as big as or even bigger than the physical asset market in the long run. In relation to this, over the last years, it was found that an increased number of cryptocurrency projects focused on building financial primitives for the NFT space. One of them are the lending platforms where users can borrow fungible tokens by collateralizing the loan with their NFT. NFT lending platforms allow users to borrow liquid assets by using their NFTs as collateral. It is possible to distinguish three types of lending pool approaches: peer to peer, peer to pool, and collateralized debt position (CDP). Starting from them, we are going to illustrate their mechanisms and provide examples of some platforms.

  1. Peer to peer: This is the main scenario that enables individuals to borrow and lend money directly from one to another. These platforms have similar behaviours. One of these is that the NFTs are locked as collateral in contract, then the lender provides a loan to the borrower within a specific time frame. Another takes into account the lending process which involves an auction on the loan’s interest rate. In general lenders who provide the lowest interest rates win the bids. The main benefit of the P2P lending is that it is perfect for covering NFTs risks. This kind of lending is characterized by combinations of high interest rates and moderate loan-to-value ratios. Platforms that use this approach are:
    • NFTfi, (2022): It is a peer-to-peer interaction among borrowers and lenders. Borrowers use their NFTs as collateral to get liquidity by the lenders that give back wETH or DAI. The system enables changes to loan settings (market conditions, life circumstances, lending/borrowing strategies, etc.). In addition, a loan can be renegotiated by the borrower or the lender. It can be triggered any time during an active loan or after the loan has expired (before the lender forecloses the loan).
    • TrustNFT‌, (2021): The main characteristic of the platform is the accurate evaluation provided by an NFT Evaluation Machine which is powered by AI and big data. The platform uses data sources from the blockchain and establishes trends in the NFT and cryptocurrency markets. It is a peer-to-peer platform for NFT-collateralized loans that enables borrowers to put up assets for loans and lenders to make offers to lend in return for interest. Once the borrower picks her NFT, the platform evaluates the maximum borrow limit for the asset(s). Following this, there is an agreement phase and then the asset is locked in a TrustNFT smart contract until the loan is completely repaid.
    • Pawnfi‌, (2020): The Pawnfi protocol proposes the concept of Non-Standard Assets (NSA) that can be fungible and non-fungible (such as NFTs, LP Tokens, Fungible tokens with liquidity, tokenized insurance, bonds, bills, derivatives, long-tail assets). Owners of these assets can draw loans against or lease their assets to gain instant usage from their illiquid tokens by placing them on Pawnfi. The system has three modes: pawn (collateral for loans), lease (lending the assets to others for passive income) and sale (listing on secondary markets for the highest bid).
    • Yawww‌, (2021): It is the first Solana's P2P automated escrow & trading service for NFTs and Solana's P2P loans marketplace. According to its protocol, every owner of Solana NFTs (borrowers) can create a loan request on Yawww using her NFT as collateral and setting loan amount, interest rate, and duration. Anybody can fund the loan with a single click, or they want to negotiate the terms.
    • Arcade‌, (2021): The system allows users to wrap multiple NFTs into one wrapped NFT that can be collateralized as a single asset. NFT price movements depend on the floor price and the value of ETH. If the value of the NFTs is lower than the borrowed coins, lenders might be at the loss, and not be able to sell the NFTs at the market price in case of late payment.
  2. Peel to pool/peer-to-protocol: This kind of system takes into account NFTs lending borrowing liquidity directly from the protocol. The main actors in this scenario are the liquidity providers and the borrowers. The former deposits into the protocol pools while the latter get liquidity after collateralizing their NFTs which are locked in the protocol’s smart digital vaults. Platforms that use this approach are:
    • NFTuloan, (2021): It is a liquidity pool provider for digital arts, domain names, collectibles, gaming cards, trading cards, virtual lands, utilities. Once the user has added her crypto wallet to the platform, she can apply for loans using her NFTs as collateral. Each loan has a fixed payback duration of 1 hour to 30 days, and the flexible interest rate which depends on the NFT. value. Moreover, NFTuloan provides an instant NFT estimation feature in order to guarantee accurate, transparent, and fair valuations of the token.
    • BendDAO, (2021): It is an NFT liquidity protocol that supports instant NFT-backed loans, Collateral Listing, and NFT Down Payment. In BendDAO, NFT holders can borrow ETH through the lending pool using NFTs as collateral instantly, while depositors provide ETH liquidity to earn interest. To minimize the losses caused by the market fluctuations, the borrower has a 24-hour liquidation protection period to pay back the loan (including interests). All the collateral are directly listed on the DBendDAO. Moreover, NFT holders can choose to get up to 40% of the floor value of the listing before it even sells. The instant liquidity is provided by the instant NFT-backed loan. The balance after deducting debt with interests will be transferred to the borrower after the deal.
    • PINE, (2022): Its lending protocol allows borrowers to borrow fungible digital tokens from lenders using non-fungible tokens as collateral. Lenders earn yield on tangible digital assets, acquire NFT assets at a discount and segregated pool structure for better market and compliance risk management. Borrowers obtain instant permissionaless loans with no back-and-forth needed with lenders and have flexibility to repay loan early or to extend loan via rollover. Every lender sets up their own segregated lending pool giving flexibility to choose the types of collateral they want to lend and set specific terms (such as loan fixed duration and interest rates). In case of default, the ownership of the NFT is transferred to the lender. Pine.loans is a platform for lenders to list the loan offers, allows NFT owners to get loans with the best conditions and guarantees enforceability of the terms of the loans.The platform offers a portal where lenders are able to manage their segregated pools, their offers and collateral repossession.
  3. Collateralized Debt Position: It is a new model for the NFT collateralized currency market which lets borrowers take out the stablecoin DAI when they collateralize ETH. It is created when collateral is locked into MakerDAO’s smart contract and the decentralized stablecoin DAI is generated. The value of the collateral placed in the CDP exceeds 150% of the generated DAI value. In undercollateralized scenario, the assets in the smart contract are sold and the funds are used to repay the generated DAI, a 13% liquidation penalty, and the stability fees applicable at the time of payment. DAI, an Ethereum-based ERC-20-compatible token, are a decentralized loan and is backed by the collateral’s value. If the borrower wants to unlock the collateral, she has to repay the DAI (and other stability fees). A platforms that uses this approach is:
    • JPEG’d, (2022): It is a decentralized lending protocol on the Ethereum blockchain that enables NFT holders to open CDPs using their NFTs as collateral. Users mint PUSd (native stablecoin of the protocol) or pETH (native Ethereum derivative of the protocol) allowing them to obtain leverage on their NFTs. JPEG'd uses a peer to protocol lending mechanism and borrowers can set up several options during the lending step. PUSd gets minted against every borrow position at a fixed borrow rate (2%) and burned upon closing of the position. All debt positions allow 32% of the collateral value to be drawn and liquidation occurs whether the debt/collateral ratio is 33% or higher.

Technical challenges

The development of the project led to face and manage several technical challenges:

  • The NFT can be any Algorand Standard Asset adhering to the standard specifications ARC-0003.
  • Economics concept to the loan (i.e. interests).
  • Any Algorand Standard Assets can be received as collateral.
  • The interest is dynamically calculated on the portion of the unpaid debt
  • The debt can be repayed at any moment, in any percentage
  • Correct handing of atomic groups is required
  • Multiple inner payments/asset transfers must be issued
  • Temporal parameters have to be calculated in terms of blockchain round. This led to some issue in Testnet already raised to the algorand repository that as been solved pushing arbitrary round number when needed.
  • The smart contract can be used multiple times in sequence without recreating the app. This is a gain for the contract owner.
  • Complex logic: we handle multiple scenarios (candle auction, offer cancel, borrower/lender unresponsive, deadlines, etc.).
  • Minumum balance management with dynamic optins.
  • 0-fee inner transactions.

Qualitative discussion of security and correctness

All the input data have been carefully checked. We defined threshold to contrain the input parameters and to avoid overflows when computing the interest. We checked the presence of manager/freeze/clawback addresses of the incoming asset to prevent unauthorized collateral acquisitions. We checked the coherence of the transactions belonging to the same transaction group (i.e., receivers, senders, amounts). We defined a state variable that simplifies state management. We performed multiple tests on the smart contract, based on 5 scenarios (file src/interact.py can be used to repeat the tests). We report a description of each scenario and the ID of some example transactions (testnet):

Scenario 0 -> App Setup. This is used to install the contract

Scenario 1 -> Loan complete flow: the lender repays the debt

  • provide_access_to_nft: CQFCJNUQQC5XLXXT7VJMXSPQMIP6OSJCHN6UEZTRD7ECRT2F4YVA
  • set_offer: TPVCOKPEAZIDXZBS3P6BFPOFVOBTL4Y6QSOCQWCXD5PGINONXLQQ
  • place_bid: S7AKDKU57VIIVWWJ4TXG6W3LEM5VWVGETQJAHIWTNYI7UGG5TVRQ
  • accept_offer: SWVNWRFORLKXPYHPT57XISNLSWDYCOIFEE2C6FT7A74N22KSSYVA
  • pay_back: CM63JNAITVZ22NDCF3URL65G3GVDOWRQ2V6YTMFMY445V6ADO4SA

Scenario 2 -> Lender calls timeout after auction period ends to give assets and algos back to their original owners

  • provide_access_to_nft: ZILENTZPUESCWQXTYAXV3TYYXRRNO6OW3QT7LRQS34D2TCNJO2RA
  • set_offer: S2FDDOZRLV77X6AK3JQB7JKGXLJUNAGPXQAFBHRW2BF5VEPAQMZQ
  • place_bid: WHZC6UBLJ5OYNLFABFXNJUI5NYVFWSDKAUWZSD6SWPQA5LPRATKQ
  • timeout: XQ6Y5NNVUTCSJJWJ7JEGZFCQN36GW5WS7ZXLO65E32CMUF2XGDXA

Scenario 3 -> Borrower calls cancel offer after mistakenly auctioning the wrong NFT

  • provide_access_to_nft: XNP74X6HITLL7YU5U6YB4G34JUF2QA5S5O3TIUXPS33JGX4U5LGA
  • set_offer: HN6QBSA2EL5QTRDUADFZYCVM25AK4OGGWWVG64EVXY24SV7AHU7Q
  • cancel_offer: LRJ5H2RM4KHKLKURS6HEJ63V7TCCGUG2KLEMAR7D7ZJWNDED25GA

Scenario 4 -> Borrower does not repay the debt in full. The lender gains the NFT

  • provide_access_to_nft: NUSYQKOAG7COUHKXE44OLIIG7V3YOIPMHVGSMFDOGCDJRHSPISNA
  • set_offer: GY7PRLWUXIXZT4Y5U5W3IEBQKWV3GJQCWYPM42RYLUAPWVD73EZQ
  • place_bid: R2YT7ZVGV7PLVHOZAERZYXYR3PQPX2DKDYGF63U2WUIBUCLCKE7A
  • accept_offer: LPPRGJQD4PTDCK7O4TGZ6LA7UKQ4CF2GUEZBRNGEIDMJ6MO7DFCA
  • pay_back: DINIDBP4QNFJVXW6GVQUGGVQKN5GT7M76HR3TZIOHEOCRDE7GSQQ
  • loan_expired: 7P7OF7IAKFNOWGBVKYJLERTPV6KMYI5HQKNWFOMHCQQCPSQRZXZQ

Scenario 5 -> The contract owner collects the fees stored in the contract

  • pay_me: 7CCPRPXI7RAI2ZDIRIM5HXMVBJE6GIFBWTOBSF3GXCDPEJ36F2ZA

Futures and business value

The NFTs loan systems are gaining increasing attention because they allow users to access to liquidity simply owning NFTs. A marketplace of this kind of contract on Algorand could allow the interoperability between the NFT marketplace and DeFi. Even more, with Algorand state proof this system may interact with other blockchain.

The developed scenario provides a profitable solution for all the users that are involved:

  • the borrower obtains her loan by using her NFT as collateral
  • the lender has an annual interest rate on the loaned Algos and obtains the NFT in case the borrower is not able to reapy the debt
  • the contract owner gains an interest rate on each loan cycle processed by the smart contract

The interaction among the users is transparent and the guaranteed by the predefined smart contract. Thus, our virtual pawnshop is not affected by the risks of phisical pownshops: the borrower is guaranteed to receive the NFT back if the loan is payed back. Moreover, the interest rate, the auction duration, and the payback period cannot be modified once the loan is taken, which guarantees transparency in the deal conditions.

There are several improvements that could be developed starting from the current scenario. One of the possible future developments is to enlarge the pool of potential lenders by considering splitting the NFT into several components. So that subsequently a community willing to lend cryptocurrency in exchange for fractional participation in the ownership of an NFT in case of default and alternatively the interest rate charged. Another potential progress is the use of stablecoin instead of classic cryptocurrencies, so that lending can be even more easily transformed into fiat currency.