Monday, May 10, 2021

Avalanche Decentralization Proposals

Mainnet AVAX Distribution with a GINI of 85.3%
Mainnet AVAX Distribution with a GINI of 85.3%

The Avalanche mainnet has a very skewed AVAX distribution, enabling the Avalanche foundation to centrally control the entire network. Below are a few proposals to amend this situation:

(1) Redesign the wallet UI to encourage people to delegate to smaller validators.

(2) Setup empty validators, and ask the Avalanche foundation to provide stakes (@kevinsekniqi’s idea): While this would not truly diminish the central power of the foundation, it’s certainly a well-intentioned first step towards decentralization.

(3a) Progressively tax large validator rewards: A governable maximum tax rate should be set for the largest possible stakes, and diminish down to zero for the smallest stakes, where the tax rate should follow a sigmoid function.

(3b) If governance of the maximum tax rate should prove too cumbersome, a governable target GINI should be determined, which would then auto-set the now dependent maximum tax rate, such that the desired target GINI would be achieved within a fixed time horizon.

(3c) Redistribute the collected tax among the validators inversely proportional to their stake size. To emphasize smaller validators even more, an exponent larger than one could be introduced.

(4) Force the foundation to burn their rewards, while allowing them to keep their stakes to ensure protection against malicious actors.

(5) Burn or re-distribute the foundation stakes, when the non-foundation-related total value locked (TVL) is judged to be high and stable enough to protect against malicious actors.

(6) Partially decouple voting power in the Avalanche protocol from the staked amount: At the moment voting weight is proportional to a validator’s stake. This could be changed to e.g. an inverse quadratic or even logarithmic relationship, where smaller validators would have relative to their stake size much more say. However, the impact on the security of such a modification should be investigated carefully.

(7) Create an Avalanche subnet named EQUALIZER, where EQUAL tokens would evenly be distributed among the currently active validators (per P-chain address): Its very existence would be a strong incentive for the foundation to take action towards more decentralization.

(8) If the foundation would not take appropriate actions to decentralize properly, simply turn the EQUALIZER subnet into a standalone network, which would stop validating the mainnet and decouple from Avalanche. It would effectively be a fork.

1 comment:

  1. heyy re point 7 - I had an extreemly similar idea which posted on Solana chat. Doing this on Subnet is a great idea. https://forums.solana.com/t/censorship-resistance-network-security-and-you/761/14 "i wanna share a premature idea i had on this.

    imagine EQ(equalizer) token. The only group who can mint it is Solana delegators. As a delegator, if the stake sizes of validator(s) you have staked to is closer to median stake size of other validators you get to mint more EQ token. If all validators stake is equal, newly minted token is distributed to every delegator equally. The value comes from investors who think EQ token communicates the right message and adds value to network’s security and censorship resistance.

    Of course as with almost every other solution, one validator can act as many. However if EQ becomes valuable, it could prevent the case where everyone be lazy and delegate to same few validators"

    ReplyDelete