Amazon EMR is the correct choice because it is a managed big data platform that provisions and manages Apache Hadoop and related cluster resources for processing large datasets such as 45 TB of nightly telemetry.

Amazon EMR can spin up and scale Hadoop or Spark clusters, automate configuration and patching, and integrate with Amazon S3 for storage so teams do not have to manage individual virtual machines and cluster orchestration themselves. This managed approach reduces operational overhead and is optimized for large batch processing jobs.

AWS Batch can run large batch workloads but it does not provide a managed Hadoop distribution or the automated Hadoop cluster lifecycle management that EMR provides.

AWS Step Functions is useful for orchestrating and sequencing tasks but it does not provision or operate Hadoop compute clusters so it cannot replace a managed Hadoop service.

Amazon EC2 supplies raw virtual machines and you can build a Hadoop cluster on top of EC2 but you would need to install, configure, scale, and maintain the cluster yourself which adds significant operational effort compared with using EMR.

Full AWS Practitioner Certification Question