A genomics research company runs a batch sequence analysis pipeline on a cluster of Linux EC2 instances. The pipeline requires a shared, high-performance file system that can deliver sub-millisecond latencies and throughput of up to 10 GB/s to ingest and process raw sequence data stored in an Amazon S3 bucket. The file system only needs to exist during the execution of the batch processing jobs, after which the processed results are written back to Amazon S3. Which combination of actions should a solutions architect take to meet these requirements? (Select TWO.)
- Create an Amazon FSx for Lustre file system using scratch storage.Answer
- Link the FSx for Lustre file system to the Amazon S3 bucket as a data repository.Answer
- CCreate an Amazon EFS file system with Provisioned Throughput and mount it to the instances.
- DProvision an Amazon EBS io2 volume with Multi-Attach enabled and mount it to all instances.
- ECreate an Amazon FSx for Windows File Server file system and mount it using the SMB protocol.
Answer
Create an Amazon FSx for Lustre file system using scratch storage, and link the FSx for Lustre file system to the Amazon S3 bucket as a data repository.
The correct architecture uses Amazon FSx for Lustre scratch storage linked to an Amazon S3 bucket. FSx for Lustre provides the high performance (sub-millisecond latencies and gigabytes-per-second throughput) required for genomics batch processing. Using scratch storage is ideal and cost-effective because data is temporary and persists in S3. The data repository link allows the EC2 instances to access and sync files directly with the S3 bucket.
Step-by-Step Solution
Key Concept
Amazon FSx for Lustre is a parallel file system natively integrated with Amazon S3. The scratch storage option is designed for temporary storage and high-performance batch processing, enabling rapid data ingestion and output synchronization back to S3.