AI will fully automate the network within five years, Rami Rahim, Juniper’s CEO, boasted at the company’s Global Summit this week.
“I firmly believe that the future, like today, will have a need for a self-driving car, that the future will revolve around a self-driving network where people literally don’t have to do anything,” he said. “It’s probably weird for people to hear the CEO of a networking company say something like that… but that’s exactly what we should wish for.”
Rahim believes that AI-driven automation is the latest phase in the evolution of computer networks, which began with the rise of TCP/IP and the Internet, accelerated by faster and more efficient silicon, and then made manageable by advances in software.
“Together, as an industry, we are sitting on an untapped gold mine of information,” Rahim said. “We as a planet are collecting data at an unprecedented rate. About 90 percent of all data in the world today was collected in the last two years alone.”
But extracting value from this data has proven difficult, he said. “There is no way a human could tap into this data and do anything useful with it themselves. You have to leave that to machines and artificial intelligence.”
And according to Rahim, AI has the potential to significantly accelerate the delivery of new services, detect and mitigate network disruptions before they impact users, and more broadly reduce the oversight required to support and manage large networks.
Juniper has spent millions of dollars on mergers and acquisitions to this end. In 2019, the company acquired Mist, an emerging wired and wireless LAN provider, not for its wireless or switch portfolio, but for its AI operational capabilities.
Since then, Mist’s AI has become a core part of Juniper’s enterprise portfolio. After the acquisition, Rahim aggressively moved to “fog” the rest of his business portfolio with AI functionality.
Alongside Mist came Marvis, Juniper’s version of a chatbot akin to Amazon Alexa or Google Assistant, to interact with the platform’s AIOps capabilities. And in late 2020, the company acquired intent-based networking startup Apstra to extend vendor-agnostic network automation to its data center networking portfolio.
“Gone are the days of vertically integrated data center network stacks,” said Manoj Leelanivas, Juniper COO. “The ability to have a multi-vendor solution where everything is flawlessly orchestrated and automated is something that Apstra brings with it.”
But while those acquisitions have helped Juniper drive automation and gain insights into the health of the network, the company still has a long way to go before AI can completely obfuscate the network, as Rahim claims.
Juniper has yet to close the loop on network automation—that is, eliminate the human element.
Juniper is not alone in its quest for AI-driven network automation. Cisco, HPE’s Aruba, Dell Tech, and Arista Networks have all announced AIOps and automation tools with the same ultimate purpose.
Cisco last week unveiled its predictive network engine, which gathers data from applications, traffic levels, protocol events and topology to predict network disruptions before they happen and potentially fix them automatically. And like Juniper, Cisco plans to expand this technology across its entire network portfolio.
Meanwhile, Aruba, which previously announced AIOps capabilities similar to those of Juniper Mist, introduced advanced automation tools with the launch of NetConductor this spring.
The platform is designed to automate micro-segmentation across large campus networks.
However, which provider will be the first to achieve a fully autonomous network remains up in the air. ®