In this BSA series – “Why AI?” – enterprise software leaders explain in their own words how artificial intelligence (AI) is having a positive impact on individuals, businesses, and organizations worldwide. In this submission, Juniper Networks Group Vice President and Chief AI Officer Bob Friday writes about how large deep models can use automation to accurately predict outcomes, such as a user’s Zoom call performance.
1. Why AI?
Networks are complex, difficult to design, deploy and operate. RF-based Wi-Fi is especially difficult to setup and troubleshoot. IT teams employed automation for decades to manage this growing complexity. AI is the final step in automation that can operate and troubleshoot on par or better than human IT domain experts. IT teams get vague trouble tickets stating the “Wi-Fi sucks!” What’s mean? Is it slow? Does it disconnect? Juniper Mist AI sheds light by learning the network with a focus on the end user experience. When a user connects to Wi-Fi, Mist AI assesses the experience—was the connection successful? Are applications slow? How’s roaming? AI visualizes the experience, identifying issues. Also, Juniper’s AI network assistant, Marvis, proactively uncovers network issues like bad cables or poor Wi-Fi coverage. Marvis finds elusive issues in less time than advanced IT experts. This reduces toil and MTTR and provides great experiences.
2. Can you give an example?
Recently we introduced a deep learning model that takes end-to-end customer Zoom data and joins that data with the network characteristics data as the network enables the Zoom conference calls, which require connection to the cloud. We ultimately trained a large deep learning model that can accurately predict a user’s Zoom call performance. This assisted a large customer whose users at a specific site were complaining about Zoom’s performance. Using the Zoom model, we were able to quickly deduce that the problem was associated with VPN distance and a VPN misconfiguration that was routing packets out of the country. This significantly reduced the amount of toil for the customer’s IT team and helped them solve their Zoom performance issue.
3. Where can we learn more?
Juniper is recognized as a leader by Gartner and your peers. Review the Gartner Magic Quadrant report for wired and wireless network infrastructure and see the Peer Insights reviews:
Join one of our weekly live demonstrations where we walk you through the Mist AI solution:
If you’re just learning about AI and would like to know how we do it, visit one of our Explainable AI (XAI) resources:
About the author:
Bob Friday is the Group Vice President and Chief AI Officer at Juniper Networks, where he leads Juniper’s cloud AI strategy. He is focused on leveraging AIOps to drive a better client-to-cloud experience.
