NFVI evolution for edge computing and cloud native
Network Functions Virtualization Infrastructure (NFVI) has been running commercially in central data centers for a couple of years. So far, the Virtual Network Functions (VNFs), or applications, running on the NFVI have been deployed with virtual machines. But NFVI is evolving to also become capable of running containers, and it’s being distributed to the edge to enable new services. This topic was brought up in a podcast with Peter Wörndle, NFV expert at Ericsson and host of the podcast, Dez Blanchfield. Here are some of the highlights from their discussion.
Service providers want to understand how applications are designed
When NFVI was introduced in the telco industry, open-source software became a reality on a wider scale. One important consequence has been that service providers have become more interested in how applications are designed. Today, it’s not enough to understand that they work as they should, and that they are interoperable. Understanding the architecture itself is getting increasingly important. For example, what components should be used, how has the software been designed and what do the development processes look like? These are the kinds of questions that have become more commonplace. As we introduce containers and microservices as part of cloud-native deployments, this interest will only continue.
Step-by-step transformation is the best approach for most service providers
Introducing NFV, cloud and edge computing are all important activities that are part of a larger transformation to improve the end-to-end network architecture flexibility and agility needed for 5G. In the podcast, discussions focused on the best way for service providers to transform. When it comes to NFV, it makes sense for most service providers to take a step-by-step approach. Start with a specific use case or an application, then move on from there, rather than implementing everything all at once.
Edge computing must make life easy for developers
There’s no single killer application for edge computing, but several 5G use cases will benefit from the introduction of this technology. To make efficient use of the edge, it must be easy for developers to work with service providers’ platforms. Developers are not interested in network topology, telco protocols and interfaces. They want to focus on designing an application that solves business needs within certain industries, such as automotive. How various network complexities impact the use case is not something developers should have to worry about. Edge computing is to a large degree about introducing new services cost efficiently, and creating simplicity for developers is a key factor in that equation.
Listen to the podcast on NFVI.