Juniper expands to accelerate adoption of intent-based networking

0
617

Juniper Networks has announced the expansion of its Contrail portfolio and new updates to the Juniper Extension Toolkit, which should bring service providers, enterprises, and cloud providers closer to the earlier promise of “The Self-Driving Network”.

With network becoming more complex, aimed at various applications, and operate across multi-cloud environments, network operators must invest in cost-effective infrastructure scaling in order to meet consumer and enterprise demands. As there was no serious push in simplifying human-to-machine interaction and complex programming languages are the only way to automate, network automation adoption is quite slow.

In order to overcome these issues, Juniper has unveiled Contrail Intent Bots and  AppFormix Analytics Bots which promise to change the way network operators interact with the network.

Juniper Bots are actually an ecosystem of new software applications focused on intent and analytics which automate workflows. Using both the Contrail and AppFormix platforms, these “bots” combine real-time analytics with automatic configuration changes across the network.

  • Contrail PeerBot automates the traditionally cumbersome process of network peering — managing multiple Border Gateway Protocol (BGP) routing and complex policy enforcement — to simplify policy enforcement and on-demand scaling.
  • Contrail TestBot, an application that’s part of Juniper’s broad effort to help network operators embrace a DevOps approach for Continuous Integration/Continuous Deployment, automates the continuous auditing of design, provisioning, and deployment changes in the network.
  • AppFormix HealthBot is a machine learning fitness and health tracker for the network that leverages AppFormix to collect real-time network data and glean insight, HealthBot translates troubleshooting, maintenance, and real-time analytics into an intuitive user experience to give network operators actionable insights into the health of the overall network.

In addition to the Contrail and AppFormix Bots, Juniper Networks has also updated its Juniper Extension Toolkit (JET) with several enhancements. These include extending the current management and control API framework of JET to the data plane, enabling developers to create applications with direct data plane access on Juniper’s vMX and MX Series 3D Universal Edge Routers. Those developers will now have access to advanced network programmability with Junos OS, providing increase control on edge routers.

“Operating a network is one of the most resource-intensive IT functions within any enterprise, cloud or service provider organization. Whether it’s ongoing maintenance or moving workloads to the cloud or deploying new services, each step has traditionally required attention to detail and countless hours of planning and execution”, said Sally Bament, vice president and service provider marketing at Juniper Networks. “Juniper Bots bring us closer to our vision for The Self-Driving Network, one that requires less human intervention and the ability to focus on more strategic business initiatives. To move at the speed of business, network operators need to spend less time precisely managing manual network knobs and levers, and more time expressing intent”, she added.

In a short talk with Networkworld.com, Bikash Koley, Juniper Network CTO, said that while the company is certainly focused on artificial intelligence (AI) deployment in the network, current AI, known as “broad AI” does not have the proper depth. While it can tell people what time their meeting is and what is the weather like, it does not have the ability to gather necessary information or be smart enough to run a network, which requires a “narrow AI”.

He added that Juniper will soon roll out more narrow AI Bots, which can be considered as a defined and narrowly focused functionality. Juniper plans to apply its R&D efforts into masking the complexity of the network allowing customers to do much more at a faster rate.

While self-driving networks are still nowhere in sight, we are slowly but steadily getting there.