These are the technologies that don't make sense as part of Network Architecture 2020. None are necessarily bad technologies, they just don't fit well into the mobile-first, application-centric world of 2020.

Though not a complete list, let's take a look at some key technologies.


Having spent my fair share of time in front of the CLI, the SNMP agent and the web GUI of each individual device in the network, it's time to say 'enough'.

For scalable, cost-efficient deployment and maintenance of network infrastructure, well thought-out network orchestration and configuration needs to be a core part of Network Architecture 2020.

It is puzzling to consider the exponential changes in network capacity and compute power over the last 20 years, and yet we are still configuring our networks largely the same way as during the 1990s.

When you think of the time needed from skilled engineers, the potential for human error and the inevitable frustration of troubleshooting, manual device configuration isn't just an annoyance - it's an active detriment to the network and to the businesses and people relying on it.

Smarter tools are needed, incorporating intent-based configuration, application flexibility and orchestration to prevent the networks of 2020 from being choked by the old way of manually configuring each device.

Imagine where home computing would be if users still had to resolve COM port conflicts and change jumpers rather than use plug-and-play USB. By and large, this is still where our networks are today from a configuration standpoint.

The potential savings in time, money and flexibility are enormous.

Standalone infrastructure

When deploying the network to support the highly-adaptable, application-driven requirements of 2020, the case for standalone infrastructure is much weaker than it has been in past years.

Co-ordination across the network and the components of service and application delivery is crucial to ensure that the most efficient use of those resources is achieved, letting network orchestration work and speeding up deployment, troubleshooting and bringing new applications and services online.

So, what are the alternatives?

Converged or hyper-converged infrastructure is one part of the solution, bringing compute, storage and even networking together tightly to allow a single orchestration platform to make changes to the system, rather than its parts.

Another part are management and operational systems that are capable of co-ordinating across multiple types and models of devices, lowering the cost and time of deployment and troubleshooting out to the network edge.

A good example of the need for this is wireless networks. Consider the difference in deploying a single WiFi access point versus deploying a dense WiFi network, indoor or outdoor. Where one access point is easy to deploy, using merely the default configuration and no wireless spectrum planning, a dense network brings in a whole new raft of challenges from spectrum planning to self-interference.

When we think of the range of technologies, wired and wireless, network and compute to support the needs of 2020, standalone infrastructure isn't just inconvenient to deploy, manage and troubleshoot - it's a detriment to the ability of the network to deliver applications, and thus value, to users and operators.


Many more are certain to join this list as Network Architecture 2020 research continues. A good place to see what could be next are the contenders.

what are the winners?

What's next?

We've seen some technologies that, as far as Network Architecture 2020 is concerned, losers.

How about some promising contenders?