We would like to:
1. Simplify the approach to indexing and reduce the number of meta-data tags necessary

• We expect this to reduce the time series cardinality and associated memory footprint

• We expect this to make the solution easier to understand

2. Remove need to maintain resource meta-data in PostgreSQL

• We expect this reduce the memory requirements of the OpenNMS system and reduce the load in PG

3. Improve visibility into time series cardinality