Agree with BGD that new deployment seems more suitable. Lower long run operating costs and gas advantages seem sufficient of a reason to support new deployment rather than upgrading. It also adheres to opt-in upgradability, which I think is a positive factor from decentralization standpoint and will also reduce execution risk.
The main negative is that some idle capital may stay stranded in the v2 protocol for a long time. For example, Aave v1 still has over $20 million in stablecoin deposits held within various OG iearn finance curve pools. The saave and aave Curve pools (using v2 stablecoins) have around $33 million in deposits, and it seems possible that a portion of this may not end up migrating over - in addition to other capital that remains in v2 due to inertia / lost keys / etc. V2 is still a great standalone protocol so it’s okay that some money will stay there, but liquidity fragmentation is not great for Aave user experience.
Some ideas for how to help minimize liquidity fragmentation and incentivize migration to v3:
- Short term (eg 3 month?) liquidity mining for key assets on v3, such as stablecoins and ETH
- Only support new asset listings on v3, not v2
- After 3-6 months from v3 launch, increase reserve factors on Aave v2 (raises protocol income while incentivizing migration)
- After 3-6 months from v3 launch, moderately reduce liquidation thresholds across Aave v2 markets (reduces management overhead, allowing Gauntlet and others to focus efforts on v3 while limiting insolvency risk)
- After 3-6 months from v3 launch, consider removing security module protection of Aave v2 Ethereum market