Service providers looking to develop more efficient network infrastructures have two key technology options. How can they decide which is best?

The way we use the internet is changing. The huge popularity of video streaming services, as well as the growing uptake of trends like wearable technology and virtual reality, are creating new expectations. Services like these typically need large bandwidths, fast speeds and low latencies. And customers want to access them flexibly, as and when they need them.

All this places new demands on service providers. And rapidly developing technologies like 5G, 4K video, and the internet of things are only going to increase the challenge they face. Cisco’s most recent Global Cloud Index Forecast predicted that annual global data centre IP traffic will reach 20.6 zettabytes (ZB) by the end of 2021 – more than three times its level of 2017.

A distributed future

Up until now, network infrastructures have never been intended to be able to cope with these demands. But service providers that want to seize opportunities for future growth are developing new ways to build their networks.

Network function virtualisation is now enabling service providers to separate the control and user planes of their network and move increasing amounts of traffic to outer layers, where it can support faster and more efficient services. Centralised data centres remain crucial for key control, security and management functions.

Keeping more traffic close to the network edge in this way will be crucial for meeting the standards of 5G and supporting the growing number of customers (like streaming services and search engines) who want to cache content in outer network layers. But it also means that service providers will need to replicate some of the traditional functions of the data centre in the outer network layers. And create connectivity across the whole network, from the core to the edge.

This is very difficult to do using existing network architectures, because it involves issues like translation across protocols and co-operation between different teams. To make a success of distributed networks, service providers will need to build their networks differently. They need network architectures that can be managed centrally, but have distributed functionality – and allow end-to-end orchestration.

Two paths to success

Cisco has developed an approach – Cloud Scale Networking – that paves the way for distributed networks by simplifying, automating and virtualising network functions. There are different ways we can do this, but they are all likely to involve at least one of two key technologies.

One of these is our Application Centric Infrastructure (ACI), which automates processes to enable networks to operate more simply and efficiently.

We know how effective ACI is, because when Cisco IT used it to automate data centres and speed up the deployment of applications network provisioning time was reduced by 58%, and capital spending by 25%.

The other technology is segment routing, a way of programming data to take more efficient routes through a network. Having pioneered the development of segment routing, we’re now using it to support service providers like Colt Technology Services make their network more efficient.

Confused? We can help you decide

So, which option should service providers go for? There’s no easy answer. Either ACI or segment routing can work. In fact, a successful approach may well combine elements of both. It all depends on your business goals and your existing infrastructure.

What we know for sure is that Cisco can support you in making the right decisions as you transform your network operations.

As leaders in this field, we’ve played a key role in developing the CORD (Central Office Re-architected as a Datacentre) open source initiative, designed to address exactly this issue.

Through work like this, we understand that creating a more distributed network infrastructure isn’t a one-off event. It’s a journey that happens in several phases, over a number of months or years. And while you’re going through this process, you need to keep your network running.

We understand this. And we understand, too, that when you’re creating a distributed network, managing security is something you need to think about at every stage, if you want to keep your systems safe.

The path to success isn’t easy. But with our expertise and technology, you can find the right way.


Find out more about how you can create a more agile and efficient network with Cisco’s data centre technology




Philippe Tubello

Manager, Systems Engineering

Global Service Provider - EMEAR