Skip to content
Riahi IT Solutions
Back to blog

May 18, 2026

AWS-Led Does Not Mean AWS-Only

How to keep AWS as the main delivery focus while still handling Azure, GCP, Databricks, and cross-cloud data integration where needed.

AWSAzureGCPDatabricks

Many data platforms are not perfectly single-cloud. A team may run most pipelines on AWS while still pulling metadata from Azure, moving data from SaaS tools, or using Databricks for Spark workloads.

The positioning

The practical approach is to keep one cloud as the delivery center of gravity. For Riahi IT Solutions, that center is AWS: S3, Glue, Lambda, Step Functions, Athena, QuickSight, Terraform, Python, SQL, and CI/CD.

Adjacent cloud work should support the delivery problem, not expand the project into a generic cloud migration. Good examples are:

  • Reading Azure API data into an AWS reporting pipeline.
  • Including GCP or Azure cost signals in a cloud efficiency dashboard.
  • Connecting Databricks outputs to AWS analytics or reporting workflows.
  • Documenting cross-cloud permissions and failure points.

AWS remains the delivery center. Azure, GCP, and Databricks are supported when they are part of the client’s real environment.

A simple decision rule

  1. Keep orchestration, ownership, monitoring, and documentation clear.
  2. Avoid turning one integration into a broad multi-cloud transformation.
  3. Put business reporting logic in the layer where it is easiest to test and maintain.
  4. Document credentials, permissions, failure paths, and support ownership.

Example scope

Goal: Add Azure cost signals to an AWS reporting pipeline.
Keep: AWS S3, Glue, Athena, QuickSight as the reporting path.
Add: Azure API extraction, validation, and documented failure handling.
Avoid: Redesigning the full cloud architecture unless the client asks for it.

The goal is not to be everywhere. The goal is to solve the data problem cleanly while respecting the systems the client already has.