Juniper Networks Inc. announced enhancements this week to its AI-Native Networking Platform, aimed at adding value to its cloud-hosted products.

These products use Mist AI to optimize user experiences and simplify operations, as well as network assistant Marvis. The new features detect and fix network, application and security issues, which enables customers to anticipate user needs. The company says this approach can lower operational costs by up to 85%.

Over the past two years, I have seen a significant shift in information technology leadership priorities. Coming out of the pandemic, chief information officers were focused on growth, remote work and customer experience initiatives. Today, security and cost-cutting trump all other priorities by a country mile, and Juniper’s ability to slash operational costs is the right use for AI today.

In a briefing ahead of the announcement, Jeff Aaron, Juniper’s group vice president of product marketing, told me that the company was launching Marvis Minis for wired, new RF spectrum analysis and predictive application assurance for wireless, plus simplified onboarding and eduroam, a Wi-Fi roaming service for educators and researchers, for NAC.

“We launched Marvis Minis in January for wireless, which now includes wired,” he said. “We’ve also added dynamic packet capture for wired, which was another unique feature we had for wireless for a while. We’ve expanded some of the Marvis actions and SLEs into wired. And we’re now adding RF spectrum capture to our troubleshooting capabilities.”

Here are the tey elements of the recent announcement:

End-user experiences

• Marvis Application Experience uses continuous learning to enhance understanding of Zoom and Teams call quality. Integrated with Zoom and Teams, it combines the Shapley data science model with ongoing user experience learning. Trained on Juniper’s data from billions of user minutes, this feature can predict call quality before a video session.
• Marvis Minis can now identify wired network issues even when users aren’t present. This artificial intelligence-native digital experience tool, which Jupiter first designed for wireless networks, has been expanded to work with wired networks. It diagnoses authentication problems without user or device involvement.
• Customizable service level expectations track critical metrics for wireless, wired, wide-area network and applications to ensure optimal user performance. The company expanded its wired SLEs, which cover throughput, switch health, and connection, to include switch bandwidth utilization, successful RADIUS server connections and switch capacity. These enhancements provide real-time monitoring and enforcement of essential wired metrics, ensuring consistent user experiences.

Troubleshooting and remediation

• AI-Native Dynamic Spectrum Capture enables “network rewind” for wireless interference. This feature provides visibility into radio frequency spectrum to help customers identify and resolve wireless interference issues more efficiently, reducing the need for site visits.
• Real-time dynamic packet capture extends to wired networks, automatically detecting issues and capturing packets in the cloud to diagnose and resolve problems without site visits.
• Marvis Actions identifies and addresses pressing issues in wired, wireless, WAN and data center environments. Existing actions for missing virtual local-area networks, port flaps, high CPU and stuck ports now include new capabilities to identify and fix misconfigured switch ports and access point loops, providing self-driving remediation and proactive suggestions for quick problem resolution.

Access control

• The Mist Access Assurance service now offers easier deployment and management through integration with Unified Endpoint Management/Mobile Device Management solutions like Jamf and Microsoft Intune. This integration ensures that devices attempting to join the network are automatically checked for compliance with corporate security standards, with specific actions taken for noncompliant devices.
• Onboarding at more than 10,000 eduroam-connected locations across more than 100 territories is now available via the Juniper Mist Access Assurance solution. It uses various authentication methods to grant access to different types of clients, including university staff and visitors via NAC 802.1x, IoT devices via self-service pre-shared keys, and guests via custom portals.

Aaron summed up the announcement and Juniper’s overall AI strategy. “In January, we put a stake in the ground,” he said. “We have AI for networking, and we have networking for AI that spans all our domains. We’re the only ones with an actual platform: common cloud, common AI.”

Final thoughts

The announcement in January that HPE is acquiring Juniper hasn’t slowed the Juniper team down. Post-acquisition news, I’ve talked to customers and channel partners who were concerned the company may pull back on R&D until the close of the deal, but that does not appear to be the case. At the MWC mobile world congress event, Chief Executive Rami Rahim was emphatic that Juniper is all-in on AI and views it as a significant sea change in networking, giving the company its biggest opportunity for enterprise share gains it has ever had.

Zeus Kerravala is a principal analyst at ZK Research, a division of Kerravala Consulting.  He wrote this article for SiliconANGLE.

Image: TheDigitalArtist/Pixabay