F5 details multi-Terabit VELOS hardware to power AI workloads
The Computer Weekly Developer Network team is at AppWorld with F5.
As part of the conference’s central newsreel, we learn that the company has announced new additions to the F5 VELOS product line.
This is a technology designed to give service providers and large enterprises a unified and customisable way to optimise modern applications required for AI, safeguard workloads and reduce downtime across single and multi-tenant application environments.
AI applications
Because AI applications are highly distributed and built with vast amounts of data traveling in complex patterns, high-performance traffic management and advanced security are critical.
Part of the F5 Application Delivery and Security Platform, VELOS is said to reduce the complexity of large-scale AI deployments by converging availability, management and security capabilities.
With the new VELOS CX1610 chassis and BX520 400-Gbps blade, the solution scales to multi-terabits of throughput, suitable for the highest performance NEBS-compliance standards and supporting the most demanding 5G Fixed Wireless Access (FWA) services.
For service providers, VELOS provides 6 Tbps of high-speed L4 throughput for FWA deployments, along with an unmatched 6.4 billion concurrent connections (400 million per blade) for exceptional scalability, and 80 million L4 connections per second for firewall security.
Both the VELOS CX1610 chassis and BX520 blade are already deployed in a production 5G network with a Tier-1 U.S. telecommunications provider.
“Making AI practical and simple for end users and practitioners is a massive undertaking, and legacy infrastructure is becoming a barrier for service providers and large enterprises seeking to implement AI-enhanced apps,” said Maggie Stringfellow, VP and GM, BIG-IP, F5. “With integrated load balancing and support for 5G workloads, VELOS optimizes performance, streamlines operations, and secures applications in support of hybrid multicloud environments, enabling apps, APIs, and large language models with the services required to unlock the true potential of AI.”
F5 VELOS offers a fully integrated and tested system, providing scalable, reliable performance for any application.
AI workloads
The promise here from F5 is that organizations will no longer need to engineer their individual application stacks with different combinations of server hardware and operating software, hypervisors and ADCs.
Moreover, VELOS dynamically allocates resources based on workload demands, optimizing traffic flow and ensuring AI models receive the data processing power they need without excess lag. VELOS addresses the increasingly complex and large data traffic volume demands of AI and data-intensive applications, maintaining low latency and consistent throughput for data flows.
As a result says F5, end users enjoy a vastly superior experience with AI applications.
