Juniper claims mist win rate against Cisco tops 80%

When it goes head to head with Cisco Meraki, Juniper Networks SVP Sujai Hajela says customers choose his company’s Mist artificial intelligence (AI)-driven WiFi technology more than 80% of the time.

“It really comes down to two companies: Cisco Meraki or Mist,” he said. “If you are looking for a real cloud, and AI-engine, our win rate is even higher than 70% to 80%.”

When asked to respond to this statement, Lawrence Huang, VP of product management at Cisco Meraki, said Cisco continues to expand its Meraki platform. “In addition, market-share reports and growth clearly highlight that Meraki is the undisputed market leader for cloud-managed networking,” he said in an email. “We are proud to continuously deliver a simple, secure and intelligent platform.”

Hajela co-founded Mist Systems and served as its CEO until Juniper bought the wireless startup for $405 million in April 2019. Last month, in its first-quarter 2020 earnings report, Juniper for the first time released Mist-specific numbers. On a call with investors, CEO Rami Rahim said standalone Mist booking exceeded a $100 million run rate on annualized bookings and grew 100% year over year in Q1.

Another question Hajela said he gets from customers is about intent-based networking. Intent-based networking allows network managers to translate their business intent by automating policy, as opposed to manually translating intent into a lot of lines of code. And although other companies including Apstra pioneered the technology, the term didn’t really take off until Cisco embraced it as its own in 2017.

“The intent-based network is now going to evolve toward the AI-driven enterprise, and the AI driven enterprise is fundamentally about user experience,” Hajela said. “Human beings can complain. If you have a problem you can complain. But let’s assume you were using a robot. Well, robots don’t complain. They just stopped working.”

It All Started With a Cisco Customer Win

Prior to launching Mist, Hajela and co-founder Bob Friday worked at Cisco. Before Hajela moved into his role of SVP of product management and strategy for Cisco’s Enterprise Networking Group, he led Cisco’s Meraki acquisition.

“We had closed a pretty big deal at Cisco,” he said. “It was with one of the largest enterprises in the world, and they had refreshed their full WiFi network with Cisco. At the celebratory dinner, the CIO walked up to Friday and said, ‘how do I know that one Monday when my first employee walks into a branch, he or she will not get a loading sign on the phone or a connectivity issue?’”

Friday didn’t have an answer. “And the CIO was like, ‘isn’t it ironic to have spent multi-million dollars to try to quantify and enrich user experience, and you just cannot tell me anything about it?”

The network architecture at the time wasn’t built to understand the user experience on the network. “It was fundamentally built to be looking at how our switch is doing, how our access point is doing, or how our router is doing,” Hajela said. “As Bob was reflecting on the conversation with me when we were both back from the dinner, we said ‘there is an amazing opportunity to really unleash the network for the next decade, which is fundamentally focused on the end user or, in today’s world, even a robot, or an Internet of Things device that network is trying to serve, rather than the switch or an access point or router.’ And that led to the fundamental thesis of Mist.”

Meet Mist, and Marvis

Mist, which got its start in 2014 as a wireless LAN company, also built an AI engine called Marvis. The founders named it after Jarvis, the digital butler in “Iron Man,” but with an M for Mist. “The AI engine became a first differentiator for us, but it wasn’t, hey AI is a cool buzzword. Now let’s figure out what problem we have to solve,” Hajela said.

A human simply couldn’t sift through the amount of data generated from an enterprise, its devices, customers, and end users to monitor in real time the end-user experience, he explained.

“The only way we could do it is by embracing math and physics,” Hajela said. “That is fundamentally what led to the data science algorithms. For us, AI was born of a necessity to handle end user experience. So we grew into the journey of AI rather than starting from AI and saying what do we need to do.”

The second differentiator, he said, is Mist’s second-generation cloud built on a distributed, microservices architecture. To support Marvis, and the huge amount of information that the AI engine processed, the company needed scalable, elastic compute. “We took inspiration from the LinkedIns, and the Twitters, and the Facebooks, and the Netflix of the world, looked at what they had done, and came out with the industry’s first, from a networking perspective, built-from-the-ground-up, modern cloud stack,” Hajela said.

Juniper’s AI-Driven Enterprise

These days Mist serves as the heart of Juniper’s AI-driven enterprise strategy. And Mavis is the brain.

Since the acquisition, Juniper has extended Mist and Marvis from wireless networks to also cover wired networks. Earlier this year, the vendor integrated Juniper’s security intelligence feed, SecIntel, with Mist, thus fusing its AI-driven enterprise and connected security strategies.

Last week the company announced new Mist-based services including proximity tracing, journey mapping, and hot zone alerting. These aim to help enterprises enable social distancing and keep employees safe as they start returning to work. They use Mist access points and cloud services in conjunction with WiFi and/or Bluetooth low energy (BLE)-enabled devices such as phones and badges.

“The Mist architecture is unique because it is the only one to combine WiFi and virtual BLE together in a unified, scalable platform,” Hajela said. “You can leverage both technologies for a variety of contact tracing use cases without having to deploy extra infrastructure hardware such as battery-powered beacons or sensors. In addition, we recently launched a premium analytics service that provides unique insight from Mist and third-party networking solutions to optimize end-user/client experiences and identify trends to assist with workplace safety.”

Mist During Pandemic Times

Hajela said he spoke with one of Juniper’s Mist customers shortly after the coronavirus became a global pandemic. This customer has upwards of 6,000 employees spread across campus and branch locations. “As soon as the pandemic hit, those employees were working from home and he said to me, ‘I’m now managing 6,000 locations because every home is a location.”

This is where Marvis “truly shines,” Hajela said, because it provides granularity into the home network to pinpoint problems — is this a Comcast router issue? Or is it coming from Zoom or Microsoft Teams or a cloud server? “They are leveraging Marvis to be able to understand each of these 6,000 connections to the home and see what’s going on.”

—SDX Central