StorageGRID 12.0: Indispensable new features for AI and modern apps 

Object storage is at the center of the rise of AI. It is used in every stage of the data pipeline: data ingestion, data prep, model training, and inferencing. Therefore, it is imperative for organizations like yours to manage this critical data securely and cost effectively. 

Available October 2, the NetApp^® StorageGRID^® 12.0 object-based storage solution provides industry-first features to manage your AI content repositories (and so much more).  

The enhancements in StorageGRID 12.0 were developed with two fundamental use cases in mind:  

• Object storage that can be used as an AI content repository at scale 
• Object storage that can be used for AI workloads with speed    

Let’s look at what’s new and how StorageGRID helps you get your business where it needs to be. 

Object storage is ideal for an AI content repository at scale. These datasets are extremely large, and customers like you often have teams of developers working on these datasets in parallel. This part of the AI workflow has been in dire need of a simple and scalable solution.

One of the long-standing axioms in AI/ML is that “changing anything changes everything.” That’s why data can be even more critical than code in the realm of AI. And while there are well-established mechanisms to version code, it’s much harder to version data. Either existing tools don’t scale, they change the data format, or they change the way that applications are expected to interact with storage.

None of it gets your AI-powered applications into production. If anything, it makes something that is already hard even harder.

StorageGRID 12.0 introduces a simple way for multiple developers to version-control their S3 buckets on StorageGRID. Your developers can make instant copies of large buckets with billions of objects and petabytes of capacity, operate on these buckets independently of each other, and reconcile changes between buckets. The creation of these S3 buckets is nearly instantaneous and takes up no additional space, dramatically simplifying your S3 workflows.

Fast access to object storage is clearly a need in the new world of AI, and NetApp is committed to helping you achieve it. As new technologies and solutions evolve, you want solutions that can work with your infrastructure today and with the least amount of developer friction.  

 To this end, StorageGRID implementation has evolved to an inner ring and an outer ring architecture:  

• The inner ring offers high speed and low latency, and it is composed of high-performance S3 caches that can deliver near-line-rate performance for working datasets. The inner ring can be connected to a specific GPU cluster. 
• The outer ring provides high capacity, high throughput, and high availability, and it can be connected to multiple GPU clusters simultaneously. 

Zooming out on this architecture’s benefits, S3 remote direct memory access (RDMA) developments are becoming a clear part of the future of object storage and AI. As S3 RDMA evolves, you want solutions that work with your existing infrastructure and that make your developers’ job easier. RDMA improves data transfer efficiencies in AI/ML environments with optimized throughput and CPU utilization, plus lower latency. 

 Another aspect to consider is caching. Caches are already extensively used in AI workflows, with multiple levels of caching—but they still present challenges:  

• Caches are complex to deploy, from an infrastructure standpoint. 
• They do not enforce a security model that aligns with the rest of the object storage infrastructure (and may not match your organization’s security needs). 
• They are not integrated, from a data consistency perspective, and your end users need to manage cache eviction and time-to-live rules.  

All these challenges are needless complexities that are only compounded with scale.    

StorageGRID 12.0 offers something better: an integrated cache that enforces security, provides a simple consistency model, and is easy to deploy. Your developers get massively accelerated performance right out of the box. This feature works across your ecosystem of S3 applications without any code or infrastructure modifications.  

  
The caching layer in StorageGRID 12.0 delivers up to 10 times the performance of current NetApp StorageGRID appliances. This performance can be further scaled up by running the caching layer on a bare-metal StorageGRID node, enabling you to customize the server to meet your specific needs. 

NetApp continues to push the boundaries of scale within a single cluster of StorageGRID. As small object workloads increase in size, StorageGRID has doubled the current published limits to support more than 600 billion objects in a single full-size StorageGRID cluster.   

And we are just getting started with what’s new in StorageGRID 12.0. Explore all the updates in StorageGRID 12.0 by reviewing the product page. And for a real-world example, dive into the story of why Banco do Brasil chose StorageGRID to solve their challenges with storage scalability, automation, and enhanced availability.