Analysts predict NFV hardware and software to be a $10+B market by 2018, yet few organizations have experience deploying NFV infrastructure at a large scale. In this talk, we will discuss lessons learned and the development that resulted as part of a specific large-scale NFV infrastructure roll-out that started in 2015.
We will discuss the OpenStack pod design for NFV that emerged after multiple iterations between Red Hat, Big Switch and Dell. This solution is currently being deployed in nationally across a tier-1 Telecom Service Provider's data centers for NFV workloads while being adapted for their central offices and mobile service centers.
The talk will cover a series of technical challenges that were overcome, and will end with some of the best practices that we learned by working together across the traditional orchestration <> networking silos.
The Telecom/NFV track is brought to you by OpenStack and OPNFV.
The discussion will focus on OpenStack for NFV at large scale, with a focus on networking / SDN topics. Design topics will include scaling out bandwidth (east/west, north/south, floating IP), automating installation, removing constraints on NFV workload placement, distributing logical routing to avoid any single point of failure and machine-assisted troubleshooting.