The Self-Driving Network: SDN On Steroids

By now the concept of Software-Defined Networking (SDN) — the emerging architecture that decouples the network control and forwarding functions enabling the network control to become directly programmable and the underlying infrastructure to be abstracted for applications and network services — has become widely known, if not widely adopted. Its market is projected to expand at a CAGR of 53% from $3.27 billion in 2016 to $42.157 billion in 2021.

But there is another SDN, the self-driving network, that is poised for an even bigger potential impact on networks, IT, and digital transformation.

According to Juniper Networks, the self-driving network™ will be able to self-discover network elements, self-configure, self-monitor and self-correct, auto-detect when a new service is needed , auto-provision to meet customer requirements and self-analyze and self-optimize —all without human intervention.

“We generally think of automation and the self-driving network as more of the ‘so-what’ of software-defined network and virtualization,” says Juniper’s Ben Baker, Director, Strategy and Marketing. “The end goal is how to run or build and manage the network in a more efficient, cost-effective way and be more flexible.”

He calls the self-driving network part of the artificial intelligence (AI) Big Bang, enabled by two recent developments: the availability of massive amounts of data needed to feed into and train deep learning systems and the availability of affordable processing power. AI will help the transition from automating routine tasks such as configuration, monitoring, and reporting to more autonomous operation. Less mundane, more creative roles will continue to require humans.

Automation is essential as network demands soar, while network resources struggle to keep pace. By 2020 there will be 4.1 billion Internet users, 26.3 billion networked devices and connections, and datacenter traffic will jump 330%. However, while automation can almost eliminate the 20% of networking tasks that take up to 80% of the staff’s time, only 16% to 30% of daily network administrative tasks have been automated.

Juniper and IDC believe machine learning systems, a subset of AI, can be trained to recognize patterns buried in the vast amounts of real-time network telemetry data. As a result, network performance issues can be discovered and remedied early, before they negatively impact customers.

Then, systems driven by policy and AI can take automatic action to change the course of events, avoiding outages affecting the user experience. “Automation will be essential — even imperative, according to Juniper — to change the current network management paradigm from one requiring constant human intervention to one where network management systems increasingly operate autonomously.”

Baker says that there are things you can do today to help yourself along the path toward a fully autonomous self-driving network, starting with his company’s automation framework that consists of the Juniper Networks’ Junos operating system, plus a suite of tools, extension kits, and programmatic interfaces. The starting point is thinking about your challenges around moving to the cloud. “The chances are it’s quite a headache for you to manage what we call the ‘cloud sprawl’.”

Juniper has an extensive portfolio — and track record — for automation, but it also sees the self-driving network as a kind of rallying cry for the industry, says Baker. “Juniper can’t do this all by ourselves. We need others to help along the way.”