V1 design

The mission of the DAO is to gather up-to-date reputation metrics of SPs in a shared data store. The initial set of the members are trusted Validators that are already collecting SP reputation data today.

Unbiased data

The Reputation Database consists of raw data and it’s not in-scope for the WG to output SP reputation scores or ranking. The Reputation Database consists of purely empirical data the Validators collect. It is up to you (aka Reputation Service) to analyze or visualize the data in a human-consumable way. This is intentionally designed for the DAO to remain neutral on what aspects of a Storage Provider service should matter more than others.

We encourage you to design your own heuristics to make sense of the data. Feel free to use any subset of the data you need for your projects.

V1 Validators

In v1, we have 10+ Validators (i.e. data sources) contributing to the reputation database, including:

filfox
filrep
filscan
slingshot_retrieval_bot
protocol_labs_retrieval_bot
new_web_group_retrieval_bot
filecoin_foundation_retrieval_bot
gravity_assist_retrieval_bot
ground_control_sp_location
lassie_bedrock
ground_control_sp_location
starboard

The full design of the data DAO is in the diagram below. The above list of Validators are represented by diamond shapes. All Validators contribute to an internal database (in black) and there is an automatic job that runs daily to sync the database replica (in purple) for public use. The reason for two databases here is to create redundancy for disaster resiliency as we iron out the details of the DAO moving forward.

Note that in v1, we do not have the tokens implemented. Therefore, all access to the database is permissionless and free.