Tech ONTAP Blogs

Next-generation Amazon FSx for NetApp ONTAP delivers higher scalability and enhanced flexibility

Doug_Chamberlain
NetApp
1,238 Views

Amazon has introduced the next-generation version of Amazon FSx for NetApp ONTAP allowing you to scale performance up to 72 GBps of throughput, that’s 18 times higher than previously. When it’s time to increase performance or storage, you can dynamically adjust the throughput of your file servers in-place or add additional more file servers to your file system, enabling you to run your most intensive and data-heavy workloads.

 

Second-gen FSx for ONTAP also showcases two new features: NVMe/TCP support for high-performance, low-latency block storage, providing a modernized, simplified, and faster alternative to iSCSI block storage. Additionally, customers can now access files almost immediately during a restore from backup, improving Recovery Time Objectives by up to 17 times for read-only workloads. More news regarding FSx for ONTAP file systems, NetApp is introducing BlueXP workload factory, a workload automation and optimization service for FSx for ONTAP. By using BlueXP workload factory, customers can more easily implement best practices into the design, setup, launch and operation of key workloads. To help users build their retrieval-augmented generation (RAG)-based generative AI applications in AWS, a joint AWS and NetApp reference architecture was created that shows developers how to use Amazon Bedrock with Amazon FSx for ONTAP to unlock new insights and capabilities.

 

New Performance Results with the second-generation FSx for ONTAP

Next-generation FSx for ONTAP systems offer a 50% increase in throughput capacity and a 160% increase in storage capacity, achieving up to 6 GBps read throughput and 512 TiB storage for a single high-availability (HA) pair. These enhancements also include over 200k IOPS per file system and 1.2 GBps write throughput. Additional gains feature a 300% higher network burst throughput (6.2 GBps vs. 1.5 GBps) and 150% higher disk burst throughput (3.1 GBps vs. 1.2 GBps). These second-gen systems enable the addition of new HA pairs and adjustment of throughput, facilitating scalable performance to meet evolving workload demands.

 

For compute-intensive workloads that require more than what is available from a single HA pair, second-generation customers can add HA pairs and adjust their throughput capacity. Second generation FSx for ONTAP file systems now allow users to scale the system from one 2-node pair to up to 24 nodes, delivering up to 72 GBps of throughput and 1 PiB of SSD storage. The ability to make these changes after the initial file system was created provides additional flexibility to optimize workload’s performance over time.

 

Furthermore, customers can adjust throughput capacity for their file systems, to incrementally scale storage (up to 1 PiB) and throughput (to as high as 72 GBps) as workload needs grow over time.  If the workload changes and needs additional or fewer resources, the customer can choose to scale all the nodes in the system up (or down) to match the latest needs of the workload. This flexibility is key for customers who are building a sustainable solution that can adjust to data-heavy workloads such as; end-to-end semiconductor design for electronic design automation, visual effects and video editing workflows, seismic data processing and simulation, and large volumes of data for analysis.

 

Customers running block workloads using the second-generation instances of FSx for ONTAP will be able to use NVMe-over-TCP as another connection protocol. As an alternative to using iSCSI, NVMe-over-TCP can offer lower latencies for some workloads, improving performance over iSCSI, providing block storage users with ways to expand workloads such as databases and virtual desktop infrastructure   to the cloud. Generally, the best performance improvements are seen in workloads with smaller block sizes with random, read-heavy workloads.

 

Near-instant access backup restores

The traditional recovery method requires users to wait for the entire data set to be restored before any of the files can be accessed. If there are files that need to be accessed quickly, using the traditional full recovery model can take a very long time to fully restore, leaving users anxiously waiting to access their files.

 

When data is recovered for second-generation FSx for ONTAP you get read access to the volume within minutes of initiating the restore, allowing you to read your entire backup data set up to 17x faster relative to backup restores on first-generation file systems. With the change to the recovery process, users are provided access to their files while a restore is in progress. The result is users do not need to wait for the entire backup to restore until they can access a file. Once the file system metadata has been restored, users can look at the file system directory structure, find the file they need to be working on, and the backup process will fully restore that file immediately, scheduling the rest to be restored later. The user no longer needs to wait for the entire restoration process to complete to have access to the file they need to use immediately.

 

Using FSx for ONTAP for implementing RAG-enabled workflows

Companies are eagerly anticipating the ability to leverage their private file data with Retrieval-Augmented Generation (RAG) to unlock valuable insights. AWS and NetApp have introduced a joint reference architecture, guiding customers on implementing RAG-enabled workflows that integrate proprietary data stored on FSx for ONTAP into their generative AI data pipelines. This solution provides guidance for a managed RAG experience for generative AI applications. By leveraging FSx for ONTAP with Amazon Bedrock, customers can use their unstructured data on file shares for RAG and enrich GenAI prompts in Bedrock with company-specific data to deliver more context-specific and accurate responses.  In addition, NetApp has introduced BlueXP workload factory to make it easier for users to deploy database, virtualization and GenAI workloads using FSx for ONTAP. These solutions offer starting points for users to build or migrate their workloads around file data stored in AWS using FSx for ONTAP.

 

Come and see the NetApp representatives at the AWS NY Summit on July 10 at the Javitz Center in New York City, booth #449 to learn more about how FSx for ONTAP might be the right solution for your cloud-based workloads. For more information on the latest FSx for ONTAP release from Amazon, please read their blog or browse the online documentation for additional details.

Public