Juniper's Mist AI navigates fog in the WAN

Juniper set its sights on further automating the wide area network, with support from its artificial intelligence engine, with today's launch of Mist 4.0.

The fourth generation of Juniper's Mist platform – Mist WAN Assurance – aims to utilize artificial intelligence (AI) to automate anomaly detection and take proactive actions in enterprises' WANs. If a client's video call isn't working, for example, the service can identify if the problem stems from the client device, application or their network connection. Juniper says Mist can proactively address the problem, often before the user is aware of the issue.

Juniper's "AI-driven WAN" provides "an AI engine making intelligent decisions in terms of policies, configurations, end-to-end service levels and troubleshooting across your LAN and your WAN as an integrated environment…an AI-hosted platform from client to cloud," says Jeff Aaron, vice president of enterprise marketing for Juniper Networks.

Mist was initially used in wireless networks, the LAN and WLAN, and now the cloud-based service provides fault resolution and security services in the WAN as well, injecting AI into Juniper's SD-WAN services. In addition, Mist WAN Assurance's AI engine utilizes telemetry data from Juniper SRX devices to provide users with the capability to customize WAN service levels.

Figure 1: (Source: Franki Chamaki on Unsplash)

"One thing that the edge of the enterprise is all about is insight into the user experience," says Aaron. "It's not about networks' experience anymore – network uptime isn't what matters – it's about user experience."

Juniper's AI-Driven Enterprise is "all about the focus on user experience, end-to-end automation, leveraging the modern cloud and using AI for support. If something goes wrong, we want to notify you before you call us," adds Aaron.

Marvis conversational interface
Mist's Virtual Network Assistant just got a little chattier with the new Marvis conversational interface, which now has natural language understanding to answer questions such as "What was wrong with Iain Morris' Microsoft Teams meeting yesterday?" While Marvis can't make Iain another cup of coffee, it can help identify network or software issues interfering with a video call.

Figure 2: Even Marvis can't fix Iain Morris's coffee problem.
(Source: Izz R on Unsplash)

Users can access Marvis via Juniper's interface or work with APIs and a third-party interface to "drill down to the root cause of the problem," says Aaron.

Juniper also updated Marvis to allow it to identify the intent behind questions, plus dialogue with users verses having a one-way conversation, says Aaron. He adds that Juniper's internal IT team uses the interface in addition to providing the service to customers.

Users can also provide feedback on Marvis' response to questions so that the AI tool continues to learn and improve. Juniper hopes that Marvis will eventually be able to answer 85% of IT users' questions, adds Aaron. Juniper customer ServiceNow has reduced trouble tickets by about 90% with Marvis – a reduction of about 100 trouble tickets a month to one or two, he says.

Mist tackles COVID-19 safety
Mist was also updated in May to address safety concerns around COVID-19. Enterprise customers can utilize Mist access points and cloud services with Wi-Fi and/or BLE-enabled devices (e.g. cell phones or employee badges) for proximity tracing, journey mapping and hot zone alerting. Mist integrates with the ServiceNow Contact Tracing app via APIs to automatically handle case management workflows for individuals who have self-reported as a potential COVID-19 risk and could have come into contact with other employees.

Juniper recently reported Q2 financial results of net revenues at $1,086.3 million, a decrease of 1% year-over-year. In addition, Juniper forecasts revenue in Q3 to reach $1,125 million, plus or minus $50 million.

— Kelsey Kusterer Ziser, Senior Editor, Light Reading