Autonomous networks will lead from the front as governments and security experts step up the fight against cyber attacks, says Adam Thomas Hartley, Network and security pre-sales technical consultant at Westcoast
We must all acknowledge the truth: IT departments are facing ever-increasing pressure from upper management due to the recent wave of system breaches and cyber attacks. The extra workload is pushing already under-staffed and under-skilled IT departments near to breaking point. To solve this issue, IT teams are realigning workloads by moving simple network and security tasks to network automation.
So what is network automation? Network automation has five main pillars: script-driven automation, automatic configuration and provisioning, automatic operation and management, high-level orchestration and policy-based networking. Using those pillars, IT organisations can offload a variety of menial tasks such as device configuration changes using scripts, or zero touch provisioning using existing hardware to configure and set up new hardware. They can even spin up and down virtual estates using SDN Policies.
The current hot questions within the industry are: ‘what is the next step on our automation journey?’, ‘what is the end goal of automation?’, ‘are autonomous networks the end goal of automation?’, and if so ‘where do humans fit in with autonomous systems doing most of the work?’
“The real power of an autonomous
network is within cybersecurity”
To answer these, we first have to understand what autonomous networks actually are. Autonomous networks are networks where we plug them in, initially set them up and away they go. They begin by learning the network’s behaviour and understanding what normal usage patterns are over time. The network then rewrites its own policies and code for network optimisation, security, auto discovering and gaining access to new resources without human interaction.
The real power of an autonomous network is within cybersecurity. When a new and unknown threat enters the network, the network then automatically and independently (without human intervention) creates new security postures within the network to respond to the new threat in real time. The network has already figured out what happened, defended against it and changed policy to stop it from happening again without any input or guidance from us humans. Gone are the days of manual crunching before resolutions can be created, let alone implemented.
Current estimations put autonomous networks within three to five years reach. This is due to the exponential growth and increasing sophistication of artificial intelligence, machine learning, deep learning and quantum computing. Progress in all these areas is driving the current innovation in network automation.
So back to the original question: ‘are autonomous networks the future?’ My thoughts are that we humans will tell the network what we want from it and the network will do it for us. This will enable the network team to transcend from admins to architects and to set out what the business needs from the network. To this extent I do believe that autonomous networks are the future.