Skip to the content.
AWS Solutions Architect – Professional
FSx
FSx For Windows File Servers
- FSx for Windows are fully managed native Windows file servers/file shares
- Designed for integration with Windows environments
- Integrates with AWS managed Directory Service or Self-Managed AD
- Resilient and highly available service. Can be deployed in single or multi-AZ within a VPC. Even in a single-AZ deployment the backend of the service uses replications to ensure that is resilient to hardware failure
- It can perform a full range of different kind of backups, including client-side and AWS-side features. On the AWS-side, it can perform automatic and on-demand backups
- FSx can be accessed over VPC peering, VPN and DX
- FSx supports de-duplication, scaling through Distributed File System (DFS), KMS at rest encryption and enforced encryption in transit
- Allows for volumes shadow copies, we can initiate previous versions for a file
- It is highly performant: 8 MB/s up to 2 GB/s throughput, 100k’s IOPS, <1ms latency
- Features:
- VSS - user-driven restores (view previous versions)
- Native file system accessible over SMB
- Uses Windows permission model
- DFS - scale-out file share structure
- Managed - no file server admin
FSx for Lustre
- File system designed for high performance workloads
- Is a managed implementation of the Lustre file system, designed for HPC - Linux clients (POSIX file system)
- Lustre is designed for machine learning, big data, financial modelling
- Can scale to 100’s GB/s throughput and offers sub millisecond latency
- Can be provisioned using 2 different deployment types:
- Persistent: provides HA in one AZ only, provides self-healing, recommended for long term data storage
- Scratch: highly optimized for short term solutions, no replication is provided
- FSx is available over VPN or DX for on-premises
- S3 repository: files are stored in S3 and they are lazily loaded into FSx for Lustre file system at first usage
- Sync changes between the file system and S3:
hsm_archive
command. The file system and the S3 bucket are not automatically in sync
- Lustre file system:
- MST - Metadata stored on Metadata Targets
- OST - Objects are stored on object storage targets, each can be up to 1.17 TiB
- Baseline performance of the file system is based on the size:
- Size: min 1.2 TiB and then we can use increments of 2.4 TiB
- Scratch: base 200 MB/s per TiB of storage
- Performance offers: 50 MB/s, 100 MB/s and 200 MB/s per TiB storage
- For both types we can burst up to 1300 MB/s per TiB using credits