[ARFC] Deployment of Aave on zkSync


[ARFC] Deployment of Aave on zkSync

Author: ACI

Date: 2024-06-10


Simple Summary

The current ARFC proposes deploying Aave V3 on zkSync to leverage the scalability and cost-efficiency of zk-Rollups. By integrating with zkSync, Aave can offer users faster and cheaper transactions while maintaining the security and decentralization of the Ethereum mainnet. This ARFC seeks governance approval for the deployment of an Aave V3 instance on the zkSync Era L2.

Motivation

zkSync is a Layer 2 scaling solution for Ethereum that utilizes zk-Rollups to batch multiple transactions off-chain and submit a single proof to the Ethereum mainnet. This approach significantly enhances transaction throughput and reduces gas fees. Deposit and borrowing activities can greatly benefit from the reduced transaction costs and increased throughput offered by zkSync.

Benefits of deploying to zkSync

  1. Reduce Gas Fees: By leveraging zkSync’s efficient transaction processing, users will incur significantly lower gas fees.
  2. Scalability: zkSync’s high throughput will allow Aave to scale effectively, accommodating more users and transactions without compromising performance.
  3. New Frontier & Opportunities: This deployment will offer the Aave DAO a new frontier for growth to explore and a new avenue path for GHO, our stablecoin.
  4. Interoperability: Due to zkSync infrastructure, an Aave V3 deployment in the zkSync Era can benefit other Aave family products, such as Lens on a segregated layer with GHO, making it a mutually beneficial relationship.

Specification

Proof of Liquidity (POL) and Deposit Commitments

  1. Aave DAO commits to redistribute to Aave users via liquidity mining, GHO secondary liquidity incentives, safety module deployments, or merit programs any airdrop provided by the zkSync ecosystem.
  2. ACI, on behalf of the Aave DAO, will coordinate any liquidity mining campaign on a zkSync Aave V3 deployment.

Risk Parameters

This section has been updated after @ChaosLabs & @LlamaRisk feedback

After Risk Service Providers feedback, the following Risk Parameters will be applied to the following assets:

USDC USDT WETH wstETH
Isolation Mode NO NO NO NO
Enable Borrow YES YES YES YES
Enable Collateral YES YES YES YES
Emode Category N/A N/A N/A N/A
Loan To Value 75% 75% 75% 71%
Liquidation Threshold 78% 78% 78% 76%
Liquidation Bonus 5% 5% 6% 7%
Reserve Factor 10% 10% 15% 15%
Liquidation Protocol Fee 10% 10% 10% 10%
Borrow Cap 900,000 2,700,000 800 50
Supply Cap 1,000,000 3,000,000 1,000 500
Debt Ceiling N/A N/A N/A N/A
uOptimal 90% 90% 80% 45%
Base 0% 0% 0% 0%
Slope1 9% 9% 3.3% 4.5%
Slope2 75% 75% 80% 80%
Stable Borrowing Disabled Disabled Disabled Disabled
Flashloanable YES YES YES YES
Siloed Borrowing NO NO NO NO
Borrowable in Isolation YES YES NO NO

Useful Links

Next Steps

  1. If consensus is reached on this ARFC, escalate this proposal to the Snapshot stage.
  2. If the ARFC snapshot outcome is YAE, coordinate with BGDLabs and Catapulta service providers for an Aave V3 instance deployment on zkSync Era.

Disclaimer

This proposal is powered by Skywards. The Aave Chan Initiative is not directly affiliated with zkSync and did not receive compensation for creating this proposal.

Copyright

Copyright and related rights waived under CC0

12 Likes

Finally a zk chain Aave can be deployed on.
Fully supporting this and eagerly waiting for the AIP.

7 Likes

Bear in mind that deploying to a zk chain like zkSync is not trivial and requires quite a bit more effort and preparation compared to an EVM equivalent chain.
You need to use their plugins for Hardhat and zkSync has a fork of Foundry in a very early alpha so it’s not the greatest tooling at the moment.

7 Likes

Hi, long time lurker, first time poster.

First off not opposed to this at all, have used zkSync and enjoy the experience but due to all wierdness going round about the recent airdrop I do have a few questions:

  1. Why was Aave given an airdrop of 8 million ZK tokens when it is not deployed there yet?

1a) Is the token allocation dependent on Aave deploying to zkSync? Will we get those tokens if we do not deploy?

1b) If 1a is True how did this allocation come about? Was there any previous communication between ACI (or potentially other delegates) and zkSync? I highly doubt zkSync would have offered 8 mill tokens just out of the hope we would deploy there.

1c) If 1b is True then why did these conversations not take place openly with negotiation input from the community (much like the potential v4 update where people expressed their concerns about the budget)

1d) If 1b is True then did ACI (or potentially other delegates) recieve any other zk token allocation for authoring and advocating for this ARFC (Not including the Aave dao allocation)?

As I said I am generally supportive of this proposal, it just looks sketchy.

1 Like

Exciting @MarcZeller thx for proposing - I’m in zkSupport of this

1 Like

We are supportive of deploying Aave V3 to zkSync for the following three reasons:

  1. Deploying on Layer2, particularly zk-rollup, is a very wise investment decision that aligns with the vision of Ethereum 2.0.
  2. Many DeFi projects have decided to deploy on zkSync, and some have already deployed on it. That means there are market opportunities on zkSync.
  3. zkSync holds Aave in high regard. As evidence of this, Aave DAO will receive the highest allocation of $830k ZK among the dapp allocations.
2 Likes

Overview

Chaos Labs supports listing USDC, USDT, WETH, and wstETH as the initial assets on the proposed zkSync Era deployment of Aave V3.
We are also considering weETH, USDe, sUSDe, GHO and ZK for future listing once these tokens have become live and liquid on zkSync.

Below is a general overview of DeFi markets on zkSync Era as well as our asset listing and risk parameter recommendations for community visibility.

Market Overview

  • TVL: $132M
    • TVL has held steady at $120M - $200M since Era’s public launch on March 24th, 2023.

  • Weekly DEX volume: $100M-$1B
    • DEX volumes peaked in March 2024 and have steadily declined since, but remain above $1B in monthly terms.

  • Among DEXes on zkSync Era, SyncSwap currently leads in TVL and weekly volume, though Maverick and PancakeSwap both maintain large and at times dominant shares.
    • SyncSwap supports three pool types:
      • Uniswap v2-style CPMM pools (”Classic”),
      • Curve-esque CSMM/CPMM hybrid pools (”Stable”),
      • and Uniswap v3-equivalent concentrated liquidity pools (”Aqua”), but with added dynamic fee functionality beyond the traditional 0.05%, 0.3%, 1.0% v3 model
    • Maverick describes itself as a Dynamic Distribution AMM, a concentrated liquidity pool model with additional position resizing customizability for LPs

  • Lending protocols and derivatives exchanges also have a presence on Era.
    • The chain’s respective category leaders are ZeroLend (an Aave V3 fork with $16M in supply TVL) and HoldStation (a perpetuals DEX that touts up to 500x leverage).
    • These two categories together draw 40K-100K weekly active addresses according to Matter Lab’s DeFi Ecosystem Dashboard.

Liquidity

The majority of liquidity for the assets in question — USDC, USDT, WETH, and wstETH — is on SyncSwap. (As mentioned before, $ZK has not yet executed its TGE.) The largest pool is USDC.e/WETH, with $23M TVL and $2M daily volume.

Screenshot 2024-06-12 at 8.36.15 AM

Pools for the newly added native USDC ($4.48M circulating) are the least liquid among the assets we recommend for listing.

As a consequence, USDC also has high slippage relative to the other assets we recommend. However, we expect its liquidity conditions to improve as Circle and its partners continue to conduct outreach to zkSync Era’s protocols.

Symbol Cost of 2% DEX Slippage Cost of 5% DEX Slippage Cost of 25% DEX Slippage
USDC $340K (to WETH) $390K (to WETH) $525K (to WETH)
USDT $1.3M (to USDC.e) $1.5M (to USDC.e) $2.5M (to USDC.e)
WETH $900K (to USDC.e) $1.6M (to USDC.e) $3.5M (to USDC.e)
wstETH $680K (to WETH) $880K (to WETH) $1.6M (to WETH)

We do not recommend listing bridged alternatives — such as USDC.e — because of the inherent risks associated with bridging.

Recommendations

Enable Borrow & Collateral

We support listing all assets as borrowable and collateralizable under reasonable limits. We do not observe a significant risk to the protocol by allowing this, as long as it is bound by a cap.

Supply Cap, Borrow Cap, and Liquidation Bonus

As described in our Metis and BNB Chain deployment recommendations, we recommend setting the Supply Cap at 2x the liquidity available under the Liquidation Bonus price impact. The Liquidation Bonus is aligned with similar assets on different deployments.

Other

We recommend aligning the Reserve Factor, Liquidation Protocol Fee, and Interest Rate curves with other deployments.

Full Recommendations

Following the above analysis, we recommend the following parameters:

USDC USDT WETH wstETH ZK
Isolation Mode NO NO NO NO YES
Enable Borrow YES YES YES YES YES
Enable Collateral YES YES YES YES YES
E-Mode Category N/A N/A ETH Correlated ETH Correlated N/A
Loan To Value 75% 75% 75% 71% 40%
Liquidation Threshold 78% 78% 78% 76% 45%
Liquidation Bonus 5% 5% 6% 7% 10%
Reserve Factor 10% 10% 15% 5% 20%
Liquidation Protocol Fee 10% 10% 10% 10% 20%
Borrow Cap 900,000 2,700,000 900 30 10,000,000
Supply Cap 1,000,000 3,000,000 1,000 300 18,000,000
Debt Ceiling N/A N/A N/A N/A $800,000
uOptimal 90% 90% 90% 45% 45%
Base 0% 0% 0% 0% 0%
Slope1 5.5% 5.5% 2.7% 4.5% 9%
Slope2 60% 60% 80% 80% 300%
Stable Borrowing Disabled Disabled Disabled Disabled Disabled
Flashloanable YES YES YES YES YES
Siloed Borrowing NO NO NO NO NO
Borrowable in Isolation YES YES NO NO NO

Capo

maxYearlyRatioGrowthPercent ratioReferenceTime MINIMUM_SNAPSHOT_DELAY
9.68% monthly 7 days

E-Mode Parameters

LTV LT Liquidation Penalty
90% 93% 1%
4 Likes

Recommendation

LlamaRisk supports the assets and parameters recommended by @ChaosLabs for deploying Aave V3 on zkSync Era: USDC (native), USDT, WETH, and wstETH.

zkSync Era shows promising trends in user adoption and developer activity. However, DEX liquidity remains relatively low and concentrated primarily on SyncSwap. The presence of official and third-party bridges is a positive factor, but the 24-hour withdrawal delay from zkSync to Ethereum mainnet could pose challenges for liquidators in volatile market conditions.

While liquidity for the newly launched native USDC is very early, we believe listing it over the bridged USDC.e token is advantageous. Native USDC avoids the inherent risks associated with bridged assets and will likely see improved liquidity as Circle and its partners continue incentivizing integrations and usage across the zkSync ecosystem.

Liquidity provisions are expected to improve rapidly over the next period, and we are likely to support increased caps and onboarding of additional assets as the ecosystem matures.

About zkSync Era

  • zkSync Era is a Layer 2 scaling solution for Ethereum that utilizes zk-rollups to batch multiple transactions off-chain and submit a single proof to the Ethereum mainnet.
  • Launched in April 2024, zkSync Era currently has a Total Value Locked (TVL) of over $750m (≈$130m excluding bridged ETH) as of June 12th, 2024.
  • The network hosts a small number of Dapps, including DEXes (SyncSwap, zkSwap, Koi, Maverick, PancakeSwap, and IziSwap) and ZeroLend, an Aave V3 fork.

image
Source: DefiLlama, June 11th, 2024

Main Tokens

The main tokens on zkSync Era and their supply as of June 12th, 2024 (source: zkSync Era Explorer):

Token Contract Supply Supply ($)
USDC.e [1] 0x3355df6D4c9C3035724Fd0e3914dE96A5a83aaf4 85,921,544 86,179,309
WETH 0x5AEa5775959fBC2557Cc8789bC1bf90A239D9a91 13,166 47,665,197
USDT 0x493257fD37EDB34451f62EDf8D2a0C418852bA4C 14,675,886 14,690,562
WBTC 0xBBeB516fb02a01611cBBE0453Fe3c580D7281011 201 13,979,599
wstETH 0x703b52F2b28fEbcB60E1372858AF5b18849FE867 991 3,587,737
DAI 0x4B9eb6c0b6ea15176BBF62841C6B2A8a398cb656 1,162,825 1,166,313

[1] Native USDC launched on zkSync (April 9th), bridged USDC is USDC.e, which remains the more widely used token.

Oracle Providers

We recommend using Chainlink price feeds for zkSync, namely:

  • ETH/USD (0.5% deviation threshold, 24h heartbeat)
  • USDC/USD (0.3% deviation threshold, 24h heartbeat)
  • USDT/USD (0.3% deviation threshold, 24h heartbeat)

Pyth Network also offers price feeds on zkSync using its “On-Demand Price Update” model. However, the methodology and data provider sources used by Pyth Network are not readily accessible.

Bridging Assets

Asset transfers are achieved through the Official zkSync Bridge and third-party bridging solutions.

Official zkSync Bridge

The Official zkSync Bridge enables asset transfers between Ethereum and zkSync Era. Its key components are:

  • Bridgehub Contract on L1: Central hub for bridges, locking L1 assets for all ZK chains. Implements registry, ETH deposits/withdrawals, and message routing.
  • State Transition Contract: Manages proof verification and data availability for ZK chains. Uses StateTransitionRegistry and deploys DiamondProxy with facets for each chain.
  • Upgrade Mechanism: Ensures all chains are updated to the latest implementation. Non-compliant chains are frozen until updated.
  • WETH Contract: Deployed from L1 WETH bridge for seamless wrapped ETH transfers.


Source: zkSync documentation

Bridging Process

  1. Deposit (Ethereum to zkSync Era):

    • Users lock tokens on Ethereum via L1 bridge contract deposit, specifying destination chain.
    • Bridgehub mints tokens on destination L2.
  2. Withdraw (zkSync Era to Ethereum):

    • Users burn tokens on source L2 via L2 bridge contract withdraw.
    • L2 bridge sends withdrawal info to Bridgehub.
    • After processing (24-hour delay), Bridgehub releases funds on Ethereum.

The bridge contract (0xD7f9f54194C633F36CCD5F3da84ad4a1c38cB2cB) was recently migrated and now holds the canonically bridged assets.

image
Source: L2Beat, June 12th, 2024

Third-party Bridges

Third-party bridges, such as Orbiter Finance, Bungee Exchange, LayerSwap & Symbiosis, expand zkSync interoperability with other L1 and L2 networks.

Liquidity

SyncSwap dominates the DEX liquidity landscape on zkSync Era. Its multiple pool types (Classic, Stable, Aqua) aim to optimize for different trading pairs and use cases. It is also worth noting that Uniswap just launched on zkSync, with other mainnet DEXs likely to follow suit.

image
Source: DeFillama, June 14th, 2024

Overall liquidity remains relatively low, especially for the newly launched native USDC. DEX liquidity is mainly concentrated around the USDC.e/ETH ($30m), WBTC-ETH ($5.1m), and USDC.e/USDT ($5.7m) pairs as of June 12th, 2024. This does not preclude liquidators from bridging assets off zkSync to complete an arbitrage loop. However, the native bridge process currently takes 24 hours, with faster options using third-party bridges, substantially increasing the risk the liquidator assumes and the complexity involved with managing hedged positions across networks or exchanges.

3 Likes

Following both community and Risk Service Providers feedback, the current proposal has been escalated to ARFC Snapshot taking into account latest risk params and assets by Risk Service Providers.

Vote will start tomorrow, we encourage everyone to participate.

1 Like

I am somewhat newer to the community, but these are all very valid questions that should be taken more seriously. While I am also supportive of the content in this proposal, it does seem that there is a lack of communication around how this came about and the involvement of certain actors, or lack thereof. I think transparency on this front would go a long way, otherwise it sets a precedent to turn a blind eye towards potential backdoor conversations.

DAO governance is in its infancy and we should all be rigorously holding all actors accountable to ensure a successful long-term future.

1 Like

After Snapshot monitoring, the current ARFC Snapshot has ended recently, reaching out both Quorum and YAE as winning option with 662K votes.

Therefore the ARFC to deploy Aave on zkSync has passed. Next step will be the implementation of an AIP for final confirmation.

2 Likes

Overview

Following the launch of the ZK token several weeks ago, we recommend adding it to the upcoming zkSync deployment. Chaos Labs supports listing ZK on Aave v3’s zkSync deployment. Below is our analysis and initial risk parameter recommendations.

ZK

Market Cap and Liquidity

When analyzing market cap and trading volumes of assets for listing, we typically look at data from the past 180 days. In ZK’s case, this is constrained to June 17th (TGE) onward. ZK’s average market cap in this span was $650 million, and its average daily trading volume was $365 million (across CeFi and DeFi). We find this market cap adequate and the trading volumes reasonable as long as they are considered when setting appropriate supply caps, a debt ceiling, and borrow caps.

Untitled

On zkSync Era, ZK’s liquidity is primarily divided between a ZK / WETH pool on SyncSwap and a ZK / WETH 0.25% pool on PancakeSwap v3. The former has $6,476,174 in TVL, while the latter has just $1,480,827 in TVL.

ZK / USD Volatility

Relative to USD, ZK’s 30-day daily annualized volatility is 139.9%, with a maximum single-day price deviation of 22.01%.

Untitled (1)

Liquidation Threshold

Considering the high volatility of ZK relative to USD, we recommend setting the LT to 45%.

Debt Ceiling

Following Chaos Labs’ Isolation Mode Methodology, we recommend an initial debt ceiling of $1M. Under the methodology for Isolation Mode, we consider two levels of probabilities for extreme price drops - Medium-High and High. We estimate the probability of an extreme price drop for ZK as High.

Supply Cap, Borrow Cap, and Liquidation Bonus

Following Chaos Labs’ approach to initial supply caps, we propose setting the Supply Cap at 2x the liquidity available under the Liquidation Penalty price impact.

Untitled (2)

Given the liquidity of ZK, we recommend a 10% Liquidation Bonus, a supply cap of 24,000,000 ZK, and a borrow cap of 10,000,000 ZK.

Recommendations

Following the above analysis, we recommend listing ZK with the following parameter settings:

Following the above analysis, we recommend the following parameters:

USDC USDT WETH wstETH ZK
Isolation Mode NO NO NO NO YES
Enable Borrow YES YES YES YES YES
Enable Collateral YES YES YES YES YES
E-Mode Category N/A N/A ETH Correlated ETH Correlated N/A
Loan To Value 75% 75% 75% 71% 40%
Liquidation Threshold 78% 78% 78% 76% 45%
Liquidation Bonus 5% 5% 6% 7% 10%
Reserve Factor 10% 10% 15% 5% 20%
Liquidation Protocol Fee 10% 10% 10% 10% 20%
Borrow Cap 900,000 2,700,000 900 30 10,000,000
Supply Cap 1,000,000 3,000,000 1,000 300 18,000,000
Debt Ceiling N/A N/A N/A N/A $800,000
uOptimal 90% 90% 90% 45% 45%
Base 0% 0% 0% 0% 0%
Slope1 5.5% 5.5% 2.7% 4.5% 9%
Slope2 60% 60% 80% 80% 300%
Stable Borrowing Disabled Disabled Disabled Disabled Disabled
Flashloanable YES YES YES YES YES
Siloed Borrowing NO NO NO NO NO
Borrowable in Isolation YES YES NO NO NO

Capo

maxYearlyRatioGrowthPercent ratioReferenceTime MINIMUM_SNAPSHOT_DELAY
9.68% monthly 7 days

E-Mode Parameters

LTV LT Liquidation Penalty
90% 93% 1%

We have updated the table in our original recommendation.

4 Likes

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.

Following the governance procedures, a proposal has been created for the DAO to activate Aave v3 ZKSync.

Voting will start in approximately 24 hours, participate :ghost:

https://vote.onaave.com/proposal/?proposalId=153

7 Likes

During some final testing procedures of the activation of the Aave v3 ZKSync pool, certain unexpected behaviour was detected. For precaution, we requested the Aave Protocol Guardian to pause the pool, until the anomaly was studied in depth.

With the pool totally empty of user funds at the moment of pause, there was no impact, and the system should be considered non-active for now.

We will update the community as soon as we progress in the research we are conducting with other Aave contributors and the ZKSync team, and define next steps for the re-activation

13 Likes

Is there any update on this?

1 Like

Hello @zackox . The research is still on-going, but we expect to do more extensive communications pretty soon.

5 Likes

For transparency with the community, we have published a report about the problem detected on activation phase on BGD. Aave v3 ZKSync activation issue report

4 Likes

should we expect a redeployment?