FSx for Lustre
Definition
AWS managed high-performance parallel file system with sub-millisecond latency and hundreds of GB/s throughput for HPC, ML training, and big data.
Use Cases
- Genomics Inc.: Processing large-scale genomic data — Genomics Inc. uses FSx for Lustre to handle thousands of genome sequencing files daily, linking it to S3 for efficient data access. (Achieved 200 GB/s read throughput, reducing processing time and costs by deleting the scratch file system after use.)
Provider Equivalents
- AWS: FSx for Lustre
- Azure: Azure Managed Lustre (preview)
- GCP: Filestore Enterprise
Frequently Asked Questions
- What's the difference between FSx for Lustre and EFS?
- FSx for Lustre is optimized for high-performance computing with low latency, while EFS is a simple, scalable NFS file system for general use.
- When should I use FSx for Lustre?
- Use FSx for Lustre when your workload is compute-intensive, requires high throughput, and involves large datasets that need fast access.
- How much does FSx for Lustre cost?
- Costs depend on the storage capacity, throughput, and deployment type (scratch or persistent). Pricing is based on usage and configuration.
Category: storage
Difficulty: advanced
Related Terms
See Also