Intel Corp. and F5 Inc. want to make AI-based inference models and workloads more secure
The longtime business partners on Wednesday said they are teaming F5’s NGINX Plus, which provides robust traffic management and security, with Intel’s OpenVINO toolkit, which optimizes and accelerates AI model inference. To boost performance, Intel IPUs offload infrastructure services from the host CPU, freeing up resources for AI model servers.
The OpenVINO toolkit simplifies the optimization of models from almost any framework to enable a write-once, deploy-anywhere approach. The toolkit is essential for developers to create scalable and efficient AI solutions with minimal code changes.
The integrated solution should also be especially of benefit for edge applications such as video analytics and IoT, where low latency and high performance are crucial. By running NGINX Plus on the Intel IPU, the solution helps ensure rapid and reliable responses, making it ideal for content delivery networks and distributed microservices deployments.
“Everything is about protecting and securing the model,” Joel Moses, who works in the office of the CTO for F5, said in an interview.
“Leveraging the cutting-edge infrastructure acceleration of Intel IPUs and the OpenVINO toolkit alongside F5 NGINX Plus can help enable enterprises to realize innovative AI inference solutions with improved simplicity, security, and performance at scale for multiple vertical markets and workloads,” Pere Monclus, chief technology officer, network and edge group, at Intel, said in a statement.
Three out of four organizations are hurdling forward with AI implementation despite gaps in data governance and security concerns, according to an F5 report, necessitating on-the-spot solutions to achieve efficient, secure workflows.
The disconnect comes amid organizational fissures in the the rapid push for AI implementation, based on a survey of more than 1,000 C-suite executives and tech workers by TE Connectivity, the global connector and sensor company.
This year, TE unearthed a vacuum between engineers and executives that the new technology will be a critical component and investment priority within the next three years.
Indeed, there is a general disconnect within many organizations over AI as “the next big thing,” according to global policy think tank Rand Corp. It estimates more than 80% of these AI projects will fail — twice the failure rate for non-AI technology-related startups.
Rand attributed the large failure rate to misalignment of goals between leadership, which often has a view of what AI can and should achieve, and the people in the corporate trenches who often lack the resources and time necessary to achieve lofty goals.
