Kubernetes was originally developed for running large-scale container based web applications, which have been typically designed with Cloud Native principles in mind enabling workloads portability. I/O intensive applications represent a category of high-end applications such as network and trading services, that require advanced networking and deterministic performance. Kubernetes has not been design to support this category of application.
Intel has been working with the industry to address this unmet gap in Kubernetes, adding to Kubernetes new networking features such as support for multiple interfaces, hardware accelerators, NUMA topology, CPU management and more.
In this webinar, we will demonstrate how Cloudify orchestrates F5 Networks’ containerized network functions running in a Kubernetes managed NFVI platform using Intel® QuickAssist Technology. The solution is based on complete decoupling between system layers to allow faster application onboarding and easier Life Cycle Management, effectively delivering “Environment as a Service” (EaaS). We will also show how to use placement policy to allow the deployment of applications across multiple clouds.