markrubens - Fotolia
When North Canton City Schools launched a new one-to-one technology initiative -- distributing more than 5,000 Chromebooks throughout the Ohio city's district -- its Wi-Fi network struggled to keep up.
"We'd have students trying to take tests, and all of a sudden, the network would just stop working. That's a problem," said John Fano, network systems administrator at North Canton City Schools (NCCS), of the district's previous Wi-Fi system, which was supplied by Cisco Meraki. "We're a Google Apps district, so everything our students do is in the cloud. Wireless is a big deal."
The school district's wireless LAN (WLAN) network must support as many as 8,000 clients at a time, connecting students and staff to mission-critical applications throughout the school day. Fano said users frequently reported problems with the former system. Its web-based management dashboard, meantime, didn't make troubleshooting easy.
"It wasn't real-time information, and it wasn't enough," said Fano, the only network specialist on the district's four-person IT team. "They had a lot of the data hidden away on the back end."
To access it, his team had to open issue-specific tickets with the vendor, which meant they first had to know, more or less, what information they were looking for.
"Trying to track some of these things down was just a nightmare," he said.
From slow WLAN to smart WLAN
When the time came to renew the system's licensing, the district's IT team decided to explore other options. In 2016, third-party reseller Technology Engineering Group introduced them to Mist Systems' cloud-managed, smart WLAN technology.
"Once we got a few demo units in here, played around with them and really saw what they could do, it was a no-brainer," Fano said. "We basically ripped our wireless network out and put a brand-new one in."
Mist, which made its debut in 2016, uses a combination of big data and machine learning to track user behavior and network performance. If problems are encountered, a cloud-based analytics engine automatically adjusts access points (APs) as needed and notifies administrators with root-cause information for immediate remediation.
The district deployed more than 300 Mist AP41 802.11ac Wave 2 APs and subscribed to the vendor's network monitoring and management service. While Fano said their former gear struggled to penetrate older school buildings' cinderblock walls and glazed tile -- with clients frequently dropping off the network without warning -- he and his team found the smart WLAN APs startlingly robust. In fact, they had to limit the devices' range to prevent network cross-talk.
"We now have the power settings dialed down to about a third of the level they came with out of the box," he said. "We had some clients that were jumping onto access points four or five rooms down the hall from them."
After some early growing pains -- the NCCS IT team spent about a month and a half tweaking network settings -- Fano said the number of end-user complaints decreased dramatically.
"It's been smooth sailing. Even in common areas, where we can see several hundred kids online at a time, the access points don't get bogged down," he said.
High-school journalism students take Wi-Fi on the road
Mist's smart WLAN also connects the district high school's Mobile Storyteller bus, which gives broadcast journalism students the opportunity to gain live, on-location video production experience.
"The whole back end of the bus [is] kind of like a server room -- racks with video gear and servers. And then there are editing stations all around the bus," said John Fano, NCCS' network systems administrator. "It's incredible."
The bus takes regular production excursions, including an annual trip to a conference in Columbus, Ohio, more than 100 miles from North Canton. When in the field, Fano said an on-board Mist AP provides a local service set identifier that allows students to produce media packages and stream live video to the school's television station.
Crucially, the cloud-based management platform allows the IT team to troubleshoot connectivity issues remotely, rather meeting the bus in the field.
AI Wi-Fi in action
When a problem does occur, Fano said Mist's AI-based analytics help his team quickly zero in on the root cause. A virtual network assistant, called Marvis, responds to natural-language questions -- no command line required.
John Fanonetwork systems administrator at NCCS
"You can say, 'Show me the issues on access point 22 over the past hour,' and the information is right there," Fano said. "It makes it really easy to troubleshoot."
Thanks to machine learning, the platform's natural language processing skills continue to improve as it interacts with network managers across Mist's customer base. With Marvis' help, Fano estimated he can now identify the root cause of a typical network problem within 15 minutes. In contrast, he said he used to spend hours -- even days, in some cases -- painstakingly tracing data and performing manual packet captures to troubleshoot issues.
Mist's smart WLAN management platform also proactively detects and flags performance anomalies, using deep learning to distinguish between normal and abnormal behavior. Network managers can then analyze and address potential issues early -- sometimes before they escalate and affect end users.
Perhaps most important, Fano said Mist's AI technology helped his team get ahead of network performance problems. The platform might alert them to a particular client struggling to connect to the WLAN, for example. Fano said they then drill into the data to see if the issue lies on the client or network side.
"We can find problems very quickly," he said. "With the very small IT departments that school districts have, any assistance that AI can provide is definitely beneficial."