As a rule, most new technologies require time to gain acceptance from a wide range of businesses and achieve awareness from the broader user community. Artificial intelligence (AI) seems to be an exception to that rule. It’s gone from a sci-fi trope to a full-blown part of daily business operations with amazing speed. Every IT and business leader I talk to has AI initiatives underway. I’ve been an industry analyst for 25 years and was in corporate IT before that. I have never seen a technology shift with this much interest.
Given the unprecedented growth in interest in AI, IT organizations are scrambling to understand how to take advantage of AI’s promise while avoiding the pitfalls that typically accompany major technological shifts. Among the vendors working to help customers succeed in the AI world is F5, a company that has lived at the intersection of applications, security, and networking since its inception.
F5, which pioneered application delivery controllers, announced this week, its next generation of ADCs – ADC 3.0—to address what it calls a “hybrid, multicloud, and AI-driven IT complexity crisis.” This is a continuation of F5 redefining the ADC to meet changing requirements. The company’s original product was a monolithic, hardware-based platform which was ideally suited to how applications were deployed. As applications became more distributed, F5 evolved its product to a software model. In the cloud era, F5 built its product to be delivered as a cloud workload or in a container form factor. Now, the ADC is evolving again to help customers successfully deploy applications in the AI era.
The Seattle-based company announced new ADCs that take a platform-based approach to AI by delivering flexibility in form factor, management capabilities, and advanced control over every aspect of application security and delivery. This makes sense since AI adoption is driving apps and infrastructure to evolve, so it’s imperative for ADCs to evolve as well.
F5 has developed a new approach for tackling AI specific application delivery and security issues. This release is well timed as most companies now have AI on the roadmap, but security remains the biggest inhibitor to deployment. The application delivery challenges haven’t been as highly publicized as security as those tend to fly under the radar. To help better understand what these are, F5 has created a top 10 list, including potential solutions for each challenge and posted it on the company’s website. This includes issues such as weak DNS practices, incomplete observability and insufficient traffic controls.
New AI-centric Architectures to Secure and Deliver Modern Apps
As mentioned earlier, every shift in applications has caused a change in application delivery and AI will do the same. It’s been well documented how AI is driving change in the compute stack and the network and ADCs can be considered part of both. Over the past decade, applications have become more dynamic and distributed, and this trend will accelerate as AI becomes more widely deployed as AI can fully automate app changes. As the nature of applications changes, the ADC must evolve to keep pace. The need for effective load balancing and other application delivery functions combined with robust security features that protect data at the application and API levels is critical as we go further into the era of AI.
New Reference Architectures to Organize AI/Machine-Learning Workflows
To address these AI-specific application delivery challenges, F5 has introduced seven core building blocks to organize AI/ML (machine-learning) workflows. These building blocks provide best practices for security, application traffic management, and platform optimization, among other critical conditions. The architecture is designed to enable customers to efficiently align their business and technical requirements for AI projects.
The reference architecture enables customers to build secure and highly scalable IT infrastructure required for AI applications, reducing risks, costs, and the need for extensive in-house design efforts. F5 has partnered with several companies to support customer AI projects, including NVIDIA, Intel, NetApp, Prompt Security, AIShield, Portkey, and OVHcloud.
Enterprise IT and security teams are dealing with unprecedented security and application delivery challenges from hybrid and multicloud-driven complexity. As AI applications make more significant and frequent requests of enterprise data stores and AI models, F5 says this requires “advanced high-performance load balancing capable of handling massive amounts of data and complex traffic patterns without creating latency.” The company also raises concerns about the appearance of new AI-focused cybersecurity threats, including “model theft, training data poisoning, and prompt injection,” as GenAI continues its rapid growth trajectory. F5 says threat actors are “working to infiltrate and hijack both the AI models and the data used to train them.” The ADC 3.0 release from F5 should help customers
F5 said it will have more announcements about ADC 3.0 at AppWorld, the company’s app and API security conference, in February 2025.
