Summary

Using metrics defined in #9939, we need to capture the storage related behaviour of Tendermint. Based on the data gathered, we need to be able to extrapolate the access patterns, amount of storage used and frequency of access of different Tendermint components. The main goal of this evaluation is to extrapolate the requirements a storage engine for Tendermint must satisfy, and prioritize them according to the data gathered. Additionally this will allow us to capture potential improvement points in terms of storage usage.

DoD:

The final output should be a:

• List of requirements for a storage backend of Tendermint in the form of an RFC including a detailed analysis of the interaction between Tendermint and the storage backend.

Additional outputs that can be derived and should be addressed in follow-up issues include:

• Refactoring of Tendermint code to optimize storage access if we observe obvious inefficiencies.

Steps that need to be taken in order to achieve the desired DoD:

• Metrics are collected when running Tendermint with different workloads (large testnets, under heavy load, for a long period of time).
• (optional) It would be very valuable if we could gather the same metrics from Tendermint users.