Post Syndicated from Jeff Barr original https://aws.amazon.com/blogs/aws/new-scale-out-file-systems-for-amazon-fsx-for-netapp-ontap/
You can now create Amazon FSx for NetApp ONTAP file systems that are up to 9x faster than even before. As is already the case with this service, the file systems are fully managed, with latency to primary storage at the sub-millisecond level and latency to the capacity pool in the tens of milliseconds. This new level of performance will allow you to bring even more of your most demanding electronic design automation (EDA), visual effects (VFX), and statistical computing workloads (to name just a few) to the cloud.
Existing FSx for ONTAP scale-up file systems are powered by a single pair of servers in an active/passive high availability (HA) configuration, and can support a maximum of 4 GBps of throughput and 192 TiB of SSD storage. The server pair can be deployed across distinct fault domains of a single Availability Zone, or in two separate Availability Zones to provide continuous availability even when an Availability Zone is unavailable.
With today’s launch you can now create scale-out FSx for ONTAP file systems that are powered by two to six HA pairs. Here are the specs for the scale-up and scale-out file systems (these are all listed as “up to” since you can specify your desired values for each one when you create your file system):
| Deployment Type |
Read Throughput |
Write Throughput |
SSD IOPS |
SSD Storage |
Availability Zones |
| Scale-up | Up to 4 GBps | Up to 1.8 GBps | Up to 160K | Up to 192 TiB | Single or Multiple |
| Scale-out | Up to 36 GBps | Up to 6.6 GBps | Up to 1.2M | Up to 1 PiB | Single |
The amount of throughput that you specify for your file system will determine the actual server configuration as follows:
| Specified Throughput |
Deployment Type |
HA Pairs |
Throughput (Per Server) |
SSD Storage (Per Server) |
SSD IOPS (Per Server) |
| 4 GBps or less | Scale-up | Single | Up to 4 GBps Read Up to 1.1 GBps Write (Single-AZ) Up to 1.8 GBps Write (Multi-AZ) |
1-192 TiB | Up to 160K |
| More than 4 GBps | Scale-out | Up to 6 | Up to 6 GBps Read Up to 1.1 GBps Write |
1-512 TiB | Up to 200K |
To learn more about your choices, visit Amazon FSx for NetApp ONTAP performance.
Creating a Scale-Out File System
I can create a scale-out file system using the AWS Management Console, AWS Command Line Interface (AWS CLI), or by writing code that calls the Amazon FSx CreateFileSystem function. I will use the console, and start by choosing Amazon FSx for NetApp ONTAP:
I choose Standard create, enter a name, select a Single-AZ deployment, and enter my desired SSD storage capacity. I can accepted the recommended throughput capacity, choose a value from the dropdown, or enter a value and see my options (more on that in a sec):
The dropdown has a helpful magical feature. If I type my desired throughput capacity it will show me one or more options:
The console will sometimes show me several options for the same amount of desired throughput capacity. Here are some guidelines to help you make a choice that will be a good fit for your workload:
Low Throughput – If you choose an option that provides 4 GBps or less, you will be running on a single HA pair. This is the simplest option to choose if you don’t need a high degree of throughput.
High Throughput and/or High Storage – Maximum throughput scales with the number of HA pairs that you provision. Also, choosing an option with more pairs will maximize your headroom for future growth in provisioned storage.
I make selections and enter values for the remaining options as usual, and click Next to review my settings. I check to make sure that I have made good choices for any of the attributes that can’t be edited after creation, and click Create file system.
I take a break to see what’s happening upstairs, feed my dog, and when I return my new 2 TB file system is ready to go:
I can increase the storage capacity and I can change provisioned IOPS as frequently as every 6 hours:
At the moment, the provisioned throughput capacity cannot be changed after creation if the file system uses more than one HA pair.
Available Now
Scale-out file systems are available in the US East (Ohio, N. Virginia), US West (Oregon), Asia Pacific (Sydney), and Europe (Ireland) Regions and you can start to create and use them today.
Learn more
— Jeff;