You must modernize on‑prem Hadoop clusters and hundreds of Airflow ETL pipelines with minimal orchestration changes. Which Google Cloud architecture fits this requirement?

The correct answer is to use Dataproc, Cloud Storage, and Cloud Composer. Dataproc is Google Cloud's managed Apache Hadoop and Spark service, making it ideal for migrating existing Hadoop clusters with minimal changes to the underlying applications. Cloud Storage provides a highly scalable and durable object storage solution that can directly replace HDFS, allowing existing applications to read and write data without significant refactoring. Cloud Composer is Google Cloud's managed Apache Airflow service, enabling you to lift and shift existing Airflow ETL pipelines with minimal orchestration changes, directly addressing the requirement. Using Bigtable is incorrect because Bigtable is a NoSQL wide-column database, not a direct replacement for general Hadoop workloads or HDFS. Converting ETL pipelines to Dataflow would require significant refactoring, violating the "minimal orchestration changes" requirement. Migrating orchestration to Cloud Data Fusion would also involve substantial changes to the existing Airflow pipelines, which is contrary to the problem statement.