This will help both simplify our database infrastructure, slightly speed up data loads, and would help
The problem is that since Redis is strictly key-value, we'll need to either disambiguate the keys (e.g. it's all good if one database uses biolink types as keys while the other uses CURIEs, since we never normalize biolink types, but we can't combine two CURIE or biolink type databases) or to configure multiple databases on the same cluster, which could make deployment a bit trickier for ITRB.
This will help both simplify our database infrastructure, slightly speed up data loads, and would help
The problem is that since Redis is strictly key-value, we'll need to either disambiguate the keys (e.g. it's all good if one database uses biolink types as keys while the other uses CURIEs, since we never normalize biolink types, but we can't combine two CURIE or biolink type databases) or to configure multiple databases on the same cluster, which could make deployment a bit trickier for ITRB.