Vega Research
Beyond proof of stake
GitHub issue: #12
Description
Both proof of work and proof of stake create a validator economics that is not necessarily in the best interest of blockchain resilience. In proof of work settings, validators tend to cluster to geographic areas with favourable conditions (primarily cheap energy). In the Bitcoin case, this has led to a concentration of far more than 50% of validator capacity in China, which endangers the point of decentralisation. In proof of stake networks, validators might cluster around available capital and concentrated with service providers. Furthermore, with the increasing tendency towards cloud computing, there is a real danger that seemingly independent validators all effectively run on a set of three or four cloud providers.
Techniques exist to design more complex voting schemes - e.g., requiring 2/3 of the votes of 2/3 of a given set of geographic domains. This does raise the question on what policy these techniques should be used to implement.
Research problem: Analyse factors in validator distribution that are likely to compromise the positive effects of decentralisation, and the are likely to occur due to economic or other factors. Assuming that technical implementation is not a problem, propose a voting policy that would minimise the identified effects.
Note: There are some constraints on what policy can be implemented in a Byzantine setting (a generalised version of the 1/3 constraint) that would be useful to take into account.