PROVIDER Β· MPP data warehouse built on PostgreSQL
Greenplum
Greenplum is the open source MPP warehouse built on PostgreSQL, where the GPORCA optimiser with columnar and external tables supports terabyte parallel analytics at near-zero learning cost for PostgreSQL teams.
3.5pts
Pros
Cons
Product Information
Detailed Notes
Overview
Greenplum arrived from SRA in 2003, was open sourced in 2015 under Pivotal's stewardship, and after 2023 VMware changes continues through the Greenplum Foundation, built on PostgreSQL as an MPP architecture with the GPORCA optimiser planning parallel queries.
For Backend & Database work the appeal is a warehouse speaking PostgreSQL, since teams fluent in PG face almost no learning curve and reuse syntax, extensions, and tooling, while external tables reach object storage and HDFS and partitioning plus columnar storage handle terabyte analytics.
Watch ongoing support, because shifting commercial sponsorship leaves enterprise procurement less clear than before with a slower community release cadence, very large petabyte workloads still trail newer cloud warehouses, and derivatives such as Cloudberry carry the lineage in some markets.
Core Strengths
- Full PostgreSQL syntax and extension compatibility gives PG teams near-zero learning cost
- GPORCA optimises cost for MPP parallel execution with strong complex joins
- Columnar storage, partitioning, and external tables query HDFS or object storage directly
- Apache 2.0 licensing brings no fees or vendor lock-in
Product Ecosystem
MPP architecture
Master and segment nodes slice data by distribution key for parallel processing with redistribution needed on expansion.
Columnar and external tables
Append-optimised columnar and partitioned tables plus external tables reading Parquet, ORC, or CSV without loading.
Tooling and ecosystem
gpfdist and gpload handle bulk loading with psql and many PostgreSQL clients compatible.
Drawbacks
- Commercial support weakened with vendor changes, requiring self-verification for procurement
- Slower community release cadence than cloud warehouses
- Expansion needs data redistribution with complex petabyte operations
- Local vendor support has shifted to derivative distributions
Use Cases
- Warehousing for PostgreSQL teams (β β β β β ) - Zero learning cost for syntax and tools
- Terabyte-scale internal reporting (β β β β ) - MPP parallel queries
- Analysing data in place on the lake (β β β β ) - External tables without loading
- Petabyte cloud-native warehousing (β β ) - Evaluate cloud warehouses
Pricing
| Plan | Price | Scope |
|---|---|---|
| Greenplum open source | $0 | Apache 2.0, self-managed |
| Derivative distributions | Per node or support subscription | Local support and patches |
| Cloud marketplace images | Cloud VM and storage cost | Self-deployed with no licence fee |
The software is free with cost in servers and staff, since an MPP cluster needs at least three machines and DBA experience, and derivative distributions bill support per node to compare against usage-based cloud warehousing.
FAQ
-
Is Greenplum still viable? Yes, the community maintains it with derivative distributions carrying the lineage, though new projects should assess long-term support sources.
-
How does it differ from single-node PostgreSQL? Syntax matches while the MPP architecture shards data across segments for terabyte analytics, whereas single-node PG suits transactional load.
-
Is expansion painful? Adding segments requires redistributing by distribution key within a maintenance window, so choosing keys well at design time reduces later work.