F5 & Scality link object storage & security for AI

F5 and Scality have expanded their partnership around a joint architecture that links F5's application delivery and security tools with Scality's S3-compatible object storage, as organisations increase investment in AI and data-intensive systems.

The tie-up focuses on integrating the F5 Application Delivery and Security Platform with Scality RING, an object storage product designed for scale-out deployments. The companies position the combined design for organisations running AI training, inference and analytics workloads across on-premises infrastructure and cloud environments.

Interest in S3-style object storage has grown as enterprises standardise their storage and access of large datasets. Many AI workflows rely on object stores for training data, model artefacts and logs, increasing the need for predictable throughput, consistent access controls and resilience across multiple sites.

F5 sees AI adding pressure on infrastructure teams already responsible for performance, availability and compliance. Both companies also cited demand for moving and protecting large volumes of data across distributed deployments.

John Maddison, chief marketing officer at F5, linked the partnership to those demands.

"AI workloads place unprecedented demands on infrastructure, particularly when it comes to moving, protecting, and accessing data at scale."

Joint architecture

The validated design integrates F5 BIG-IP with Scality RING and presents it as an S3 environment spanning multiple nodes and sites. BIG-IP sits in front of the object storage layer to handle traffic.

F5 routes and load-balances S3 traffic across storage nodes and sites using DNS controls and traffic management functions. The goal is to remove single points of failure and maintain service continuity if individual components or sites degrade.

Security is the other part of the integration. The companies describe the design as combining Scality RING CORE5 cyber-resiliency features with F5 security functions, including a web application firewall, DDoS protection and policy-based access controls. They also cited TLS offload and hardware-accelerated cryptography for handling encrypted traffic.

Scality says RING provides petabyte-scale object storage with durability and self-healing mechanisms. It is designed for multi-site deployments and long-term retention, common requirements in regulated industries and data-heavy AI programmes.

Hybrid deployments

The companies are targeting organisations running hybrid and multicloud architectures. Many enterprises train models on a single infrastructure and deploy inference services closer to end users or applications, while data governance policies often require strict control over where data resides and how it moves.

In these environments, object storage may span multiple sites, raising questions around latency, replication, access control and operational consistency. The F5 and Scality design places a traffic management and security layer in front of the object store, rather than leaving applications to manage connections to individual storage nodes.

Multi-site data protection and disaster recovery are core use cases, along with long-term retention of data such as training datasets, security logs and archived analytics.

Operational impact

The joint design is positioned as simplifying operations through centralised management and flexible deployment models. The companies also tied it to lower complexity and total cost of ownership as data volumes grow, but did not disclose pricing or customer adoption figures.

Paul Speciale, chief marketing officer at Scality, framed the partnership around the scale and resilience requirements of AI programmes.

"Enterprises adopting AI need data infrastructure that grows seamlessly without compromising cyber resilience, data protection, or uptime."

The companies said they will continue to develop and validate the joint architecture for organisations that use S3-compatible storage as a core layer for AI and analytics workloads across distributed environments.