1.8 KiB
Doc Store PG
Mongo-compatible document workloads with search on PostgreSQL with CloudNativePG
A practical research program to test wether document storage can become a composable PostgreSQL capability -- and where MongoDB still wins.
Goal: turn architecture debate into reproducible evidence.
Background
Initial Project for Research Tech Lead at Helix Forge.
Reference Stack: one PostgresSQL substrate, multiple document surfaces
Measured compatibility, not assumed compatibility
Mongo API Surface
Drivers connect through DocumentDB GW or FerretDB
PostgreSQL document core
BSON / JSONB, document APIs, indexes, SQL hybrid queries
CloudNativePG operations
HA lifecycle, extensions, databases, backup and restore
The question
Can we preserve MongoDB developer ergonomics while gaining PostgreSQL's relational model, extension ecosystem, and CloudNativePG operating model?
Measure the ladder: MongoDB -> PostgreSQL JSONB -> DocumentDB APIs -> gateways
Tech Lead Mission
Turn the hypothesis into a working lab, a benchmark discipline, and
decision-grade evidence.
Build: reference deployments Benchmark: workload ladder Stress-test: tenant isolation ladder Decide PostgreSQL vs. MongoDB trade-offs
First milestone
Repeatable lab, first compatibility matrix, tenant-isolation result, written recommendation.
Why this is attractive for a lead
It is small enough to build, deep enough to matter, and concrete enough to prove.
The outcome is not a benchmark blog post -- it is an operating model for document storage inside a PostgreSQL-first platform.
Follow Up Potential
- Extend scope to include search capablility with PostgreSQL as an alternative to ElasticSearch.
- Utilize DocStorePG as the backend technology for markitect and derivative projects including coulomb.social.
xxx