05. Databricks vs AWS Native Services
Databricks on AWS should not be treated as a replacement for every AWS service. A mature architecture usually combines Databricks with AWS infrastructure primitives.
Service mapping
| Requirement | AWS service | Databricks role |
|---|---|---|
| Object storage | Amazon S3 | Lakehouse storage foundation |
| Identity | IAM / IAM Identity Center | Cloud identity integration |
| ETL | AWS Glue | Alternative/complement to pipelines |
| Warehouse | Amazon Redshift | Alternative/complement for SQL |
| Streaming | Kinesis / MSK | Source/transport |
| Functions | Lambda | Event-driven application logic |
| Events | EventBridge | Event routing |
| ML | SageMaker | Alternative/complement to Databricks ML |
Databricks + S3
AWS Foundation
S3 | IAM | VPC | KMS
│
▼
Databricks
│
┌─────┼────────┐
▼ ▼ ▼
ETL SQL ML/AI
Databricks vs Glue
Do not describe Databricks as “Glue with a better UI.” Compare ETL complexity, Spark requirements, developer experience, governance, SQL analytics, ML/AI, orchestration, operational model, and cost.
Architecture principle
Use AWS for foundational cloud capabilities where AWS is the natural control plane, and use Databricks where the data and AI workload benefits from the Databricks platform.
Next
Continue with the next topic in 01. Databricks Foundations.