iTranslated by AI

The content below is an AI-generated translation. This is an experimental feature, and may contain errors. View original article
📖

re:Invent 2025: Analyze File Data Directly with AI Using S3 Access Points for FSx for NetApp ONTAP

に公開
2025/12/10

Introduction

By transcribing various overseas lectures into Japanese articles, we aim to make hidden, high-quality information more accessible. This time, we are focusing on the following presentation for this project!

For transcription articles of re:Invent 2025, please check the information compiled in this Spreadsheet.

📖re:Invent 2025: AWS re:Invent 2025 - Data innovation: Amazon S3 Access Points for FSx for NetApp ONTAP (STG219)

In this video, Ben Littman, Senior Specialist Solutions Architect at AWS, explains the new features of FSx for NetApp ONTAP and S3 Access Points. It is a revolutionary feature that makes hundreds of petabytes of file server data held by enterprises S3-compatible in real-time without the need for copying or moving. FSx for NetApp ONTAP can be provisioned in about 30 minutes, and SQL databases can be compressed by 80 to 90 percent through deduplication and compression. With S3 Access Points, you can connect S3 compatibility directly to NFS/SMB share data and seamlessly integrate with AI and analytics services such as Athena, EMR, Bedrock, and QuickSight. The video introduces practical examples, such as searching for compounds in pharmaceutical research, where value can be instantly extracted from petabyte-scale data using natural language queries.

https://www.youtube.com/watch?v=yg4QclLNOps
Note: This article is automatically generated while maintaining the content of the existing lecture as much as possible. Please be aware that there may be typos or incorrect information.

Main Content

Thumbnail 0

S3 Access Points for FSx for NetApp ONTAP: New Feature to Utilize File Server Data for AI and Analytics

Thank you all for joining. By the way, I am not Luke. My name is Ben Littman. As a Senior Specialist Solutions Architect, I support the storage team. I mainly help customers migrate workloads, usually from VMware platforms to AWS, and ultimately, I end up talking a lot about file servers and storage. That's because they make up a large part of our enterprise IT story. Well, today I'm going to talk about S3 Access Points. It's truly exciting stuff.

Thumbnail 30

Thumbnail 50

I'll jump right into that, but let's take a step back. Let's broaden our perspective. Why are we talking about data? Data is valuable to your company. This is exactly how you drive innovation, especially in fields like AI, machine learning, and analytics. We have massive amounts of data on file servers. Almost everyone has file servers, and we're talking exabytes of data. It's not uncommon to have hundreds of petabytes of data on SMB and NFS shares. The problem is, how do you extract value from the data you already have? You could make copies, or put it into S3 and keep two copies maintained and replicated, but there has to be a better way.

Thumbnail 70

Thumbnail 90

Thumbnail 100

Most of this is unstructured data: reports, video files, CAD files, all kinds of data. So, how do we make decisions based on this data? How do we perform analytics on this data? That's what I'm talking about today. However, when you're using file server data, you're not actually making the most of it. It's just sitting there, with teams accessing user shares, or being used and interacted with in business processes. It's not actually adding that much value.

Now, let's talk about FSx for NetApp ONTAP. For those who don't know, it's a fully managed AWS service and the only service with full functional compatibility with on-premises NetApp. You can deploy FSx for NetApp ONTAP on AWS. It takes about 30 minutes to set up, and then you have a full-featured NetApp that can handle NFS shares, SMB shares, iSCSI, and NVMe over TCP. You can replicate with SnapMirror, and all those fun features NetApp users love—snapshots, clones, deduplication, compression, all that enterprise power—are available on AWS. So, when we talk about migration, I ask, "What kind of storage do you have?" Let's move to NetApp so you don't have to manage your own file servers anymore. Nobody wants to manage file servers, right? NetApp handles that for you.

Thumbnail 150

I mentioned provisioning; you can get this up and running in 30 minutes. You can do all kinds of interesting things. For example, you can use it as a data or log volume for SQL databases. Deduplication and compression help optimize that, and it's fully managed. If there's an issue, call AWS and talk about your EC2 instances; we can support ONTAP as well. Cost is also a major point. On EC2, we often use EBS, Elastic Block Store, and there's nothing wrong with that. However, FSx for NetApp ONTAP offers, for instance, deduplication, compression, and thin provisioning. The SQL database I mentioned earlier can be compressed by 80 to 90 percent. So, this is also a way to save costs and take advantage of benefits like space efficiency, compression, compaction, and deduplication.

You can reduce complexity. Think about EBS volumes. You might have hundreds or thousands of instances, each with several volumes. With ONTAP, it's essentially one array. All volumes are there. Storage Virtual Machines can export these as iSCSI LUNs. Everything is in one place, and you only manage one thing. This fosters innovation. You can combine the power of NetApp and AWS.

Thumbnail 230

However, what I'm going to talk about next is how to perform this AI and analytics. How do you add value to this data? Because that's what leadership is looking for. That's what the business really wants. The files themselves aren't that interesting, but the analytics are. You might have decades of reports and briefings, and you'll want to ask questions via QuickSight. "Tell me about this data." Traditionally, you could move this to S3, then use Athena, Glue, and these services to crawl, read, and utilize the data. But there are all kinds of use cases where ONTAP can help. For example, backup and disaster recovery. You can use SnapMirror from one region to another. Since you are transferring data after deduplication and compression, it's a very efficient process.

Thumbnail 280

This is what I'm talking about today. This is the portfolio I'm presenting to help you add value to your data. Under analytics, we have EMR, Lake Formation, and Athena, which can crawl these files or run queries using Redshift. And you've heard about AWS's AI service, Bedrock. I also mentioned QuickSight. We're going to surround all this data with these tools and add value. That's the name of the game here.

Thumbnail 310

This is what we are announcing: S3 Access Points. So, what is this? You have a file server and you have a volume in FSx for ONTAP. All you need to do is provision an access point and say it provides access to that volume. No copying is required, no data movement is required, and suddenly, this data is fully S3-compatible. At the same time, it remains file data. This means that instead of creating copies and refreshing them periodically, you can access the data where it lives with S3 compatibility in real-time, regardless of whether you're using NFS, SMB, or any other protocol.

Thumbnail 350

This is a real game-changer. If you've been asking, "I have all these file servers, how do I make AI interact with them? How do I add value to this data?" or "How do I do more with my data?" With S3 Access Points, you can leverage all this data. It might be an SAP HANA dataset for genomics. All of this becomes S3-compatible instantly. You might have trading data and want to create reports and dashboards for your executives. Seamless integration with all analytics, AI, and compute services is possible. That's because these are designed to interact with S3. In other words, you're just connecting S3 compatibility directly to your FSx file system.

Thumbnail 390

Thumbnail 400

Thumbnail 410

So, let's look a little deeper. Let's see what else you can do to maximize this potential using this new feature. I'll show you a quick demo. This is the QuickSight interface. You just need to ask a question about your data. Let's say you have petabytes of data. It might be documents, or research data. You'll also have organizational knowledge. And that's the problem, isn't it? How do you access it? How do you add value? By simply connecting S3 Access Points to it, everything can communicate with it immediately, as if it had been copied to S3.

Thumbnail 430

Thumbnail 450

You might be using QuickSight, Bedrock Knowledge Bases, or a RAG knowledge base. Then, you can have an AI assistant that knows about all these documents and you can ask it questions. It can instantly search and run queries, letting the AI model know everything about your data. And all you have to do is throw a natural language query at it. You don't want to migrate data. You don't need ETL pipelines moving data back and forth. If you do that, you'll be paying twice. Once a file copy is made, it will always fall out of sync. This just connects S3 compatibility directly while keeping the data in FSx for ONTAP.

Thumbnail 470

Thumbnail 480

This adds value. It helps you transform underutilized file data. Users are using file shares to interact with documents. Now you can add AI to that, and instantly do much more with that data. This is one of the biggest use cases. If you're migrating to AWS and have file servers, instead of placing them on EC2 using EBS volumes, you just copy the file data directly to FSx for ONTAP and connect the S3 Access Points directly to it.

Thumbnail 490

Thumbnail 500

Imagine one of the use cases we talked about. Suppose a pharmaceutical researcher asks, "What were the promising compounds in the cardiovascular trials from 2015 to 2020?" That's a great natural language query, right? And the AI model can crawl the data and provide the answer to that exact query in real-time. You don't need to write dashboards or build this manually. That's it. Do you have any questions? This is a really great feature. It works by just connecting it, and it's quite simple. Are there any of you here who use FSx for ONTAP? Anyone? Great, perfect. Then you're already used to it. It's better than managing file servers, right?


Note: This article is automatically generated using Amazon Bedrock while maintaining the information from the original video as much as possible.

Discussion