Architecture of the week · Week 41, 2026
BigQuery Lakehouse
A zoned data lakehouse: raw and curated Cloud Storage zones feeding BigQuery through Data Fusion and Dataflow, with Dataproc for ad-hoc Spark.
Google Cloud9 resourcesData & Analytics · advanced
Why this architecture works
- Raw and curated zones separate immutable landings from validated, consumable data.
- Data Fusion provides governed, low-code ELT with built-in lineage tracking.
- BigQuery external plus native tables let you query the lake without duplicating everything.
- CMEK via Cloud KMS keeps encryption keys under your control for compliance.
- Dataproc spins up ephemeral Spark clusters against the curated zone, avoiding idle cluster cost.
Coming up
In 1 week
Hybrid Connectivity via HA VPN
Google Cloud · 8 resources
In 2 weeks
Document AI Pipeline
Google Cloud · 9 resources
In 3 weeks
Video Upload & Streaming
AWS · 9 resources
Previously featured
A new architecture every week.
Every pick is a real, editable diagram — open it, change it, and generate the Terraform that builds it.