In Pursuit of the Self-Driving Network™, Juniper Unleashes Bots Upon the Network
Dec 7, 2017
Think about how you interact with a digital assistant on your smartphone – whether it’s Siri, Alexa, Cortana or Google Assistant – and trying to play Rolling Stones on your new wireless speaker while you’re cooking dinner. Instead of manually opening the music app, searching for that particular artist and song, pressing play and double-checking to ensure you’re playing it through the right speaker, you can now just tell your device to “play Rolling Stones in the kitchen,” and your device should do the rest.
Now apply that to the network. With the advent of multi-cloud, distributed applications and service delivery, operating a network has become increasingly valuable – and also very complex. Imagine if you could just tell your network the outcome you wanted without having to worry about all of the steps along the way.
Today we’re solving that. Juniper Bots, an ecosystem of new software applications that automate operational workflows and break the cycle of complex machine-focused automation tools, were announced today. With these software applications, network operators get an “easy button” to a growing number of complex network challenges. Juniper Bots will accelerate the adoption of network automation by leveraging a natural human-to-machine interactive experience.
Network operators have traditionally interacted with the network through basic command-line interfaces, or application programming interfaces, which present information in a manner that makes it easy for machines to process and interpret. The problem is this deference to machine-based languages has given rise to highly tailored implementations that only networking experts with years of specific programming expertise can really operate. And it still takes them too much time. This results in specific managed service offerings (such as an IP VPN) or customized tools designed to create a network path that matches an application’s requirements regarding jitter, latency or capacity.
While the idea of defining intent or policy has become the preferred method for dealing with network operations, the underlying technology that makes it possible lags behind in its ability to simplify human-to-machine interaction. As engineers build tools for other engineers, a culture has emerged where humans interact with machines on the machine’s terms. And as gifted as any given engineer can be, human error still permeates. To truly advance the way humans interact with machines and advance the adoption of automation, we believe a significant change needs to take place – one where the interaction with machines is more intuitive and natural for us humans.
Juniper Bots helps usher in this new instinctive experience by translating intended business outcomes, written or spoken in straightforward natural language, into actionable automated workflows. These bots are created with six foundational principles in mind: aligning to a higher layer business intent; contextual automation; historical and predictive multidimensional views; machine learning; deterministic assurance; and continuous resource optimization.
The grand automation challenge that Kireeti Kompella issued to the industry is to cooperatively innovate on foundational technology building blocks (programmability, telemetry, intent-based networking, multidimensional views, and machine learning) to bring about a truly Self-Driving Network™. This vision of enabling networks that are autonomous and programmed to carry out an organization’s intentions and that predict performance problems without human interaction will happen thanks to the innovation Juniper Bots provides.