Dartmouth College and the AI Network of the Future

Dartmouth’s implementation of an AI-driven wireless network from Mist has revolutionized the wireless network experience for both the IT department and the college’s network users, enabling fast and reliable Wi-Fi.

By Mitchel Davis, Vice President and Chief Information Officer

Universities are under intense pressure to meet ever-expanding student wireless expectations for easy, fast, and reliable access to the applications they need, no matter where there are on campus. At Dartmouth, this effort to create the best service experience possible is getting a huge assist from artificial intelligence (AI).

The field of AI research was founded at a workshop held on the Dartmouth College campus during the summer of 1956. Those who attended would become the leaders of AI research for decades, and many of them predicted that a machine as intelligent as a human being would exist in no more than a generation.

Given that history, an AI-driven network is a natural fit for the college, and it has resulted in quicker connections, fewer dropouts, less downtime and, if something does go wrong, a much easier and quicker process to determine the problem’s cause and repair it. The benefits extend across both wireless and wired networks, which is important to deliver a solid online experience for our users, customers, clients, and partners.

AI capabilities combined with Bluetooth Low Energy (BLE) drive our campus vision of the future, where smartphones are location-aware personal assistants that provide directions, guide museum tours, and allow automated individual classroom configuration along with IP-driven smart buildings and other services. We are treating IT modernization as a key strategic investment that will lead to entrepreneurial research and teaching opportunity, along with streamlined processes and cost savings.

Many universities (and organizations in general) have wireless networks built on obsolete and aging architectures that struggle to keep up with growing demand, are hard to manage and troubleshoot, and fail to meet the extreme demands of today’s always-connected users.

For example, today’s students—the first digitally native generation that has never known a world without the internet—expect fast, pervasive, reliable wireless service in their university experience as much as an engaging curriculum, good professors, and a vibrant social life. Woe to the institution that can’t provide it in the classroom, the dorm, the library, or wherever.

Consider Dartmouth’s Wi-Fi network. In fall 2011, we had 4,000 students using it. Today, with Ethernet ports as easy to find on campus as desktop computers, that number has grown to 22,000. Most laptops don’t even have a built-in wired port, and 75% of our students have never used one.

Dartmouth’s Wi-Fi network is many things to many people. Most of our undergraduates (90%) live on campus in Hanover, New Hampshire, and they expect unfettered access, whether they’re emailing a paper to a professor or streaming a movie on Netflix. Professors want to be able to pair their iPad to an AppleTV in the classroom and pull up visual aids for the day’s lecture, and administrative staff members need dependable service to do their work.

As the number of users and devices soared, so did the challenges we faced managing the network. Service levels were inconsistent. Complaints to the help desk from students and staff were high. The overwhelmed IT staff were forever in a game of whack-a-mole, poring through log data from different servers and routers in a frantic search for what went wrong, only to see another issue pop up once they figured it out.

We were in dire straits because like many universities and other large organizations, we were constrained by network architectures designed more than 10 years ago, before the advent of the smartphone, cloud computing, AI, BLE, and other innovations.

It was a setup that gave us little visibility into how users were experiencing the network, which constantly put us in a reactive mode rather than helping us proactively gain insights into the network’s operation and identify problems as soon as or even before they affected users. With new AI-powered wireless technology from Mist, we acquired the capabilities to change everything. The Mist platform collects 150 pieces of information every two seconds about what the wireless user is experiencing, analyzes the data in the cloud using machine learning, and instantly returns detailed information about where the problem originated and why.

For the first time, we’re able to establish, measure, and enforce defined service levels. For example, we can state that it should take no more than two seconds for a smartphone user in Baker-Berry Library to get on Wi-Fi—and know right away that it took someone six seconds and whether it was the device or some specific part of the network that failed.

Even better, the technology anticipates issues before they occur, adding an unprecedented “self-healing” quality to wireless network performance. The technology we purchased even offers a Siri-like “virtual network assistant” that uses natural language processing (NLP) to provide network administrators with answers to questions such as “How are the Wi-Fi access points in Goldstein Hall performing?” Complaints to the help desk have dropped sharply, and there’s a strong feeling that everyone across campus is enjoying a common experience across wired and wireless.

We’re also excited about where the new technology may take us in the future, with its BLE features that enhance the mobile experience beyond basic Wi-Fi connectivity. The BLE capabilities in everyone’s smartphones can interact with wireless beacons to, say, give a new student turn-by-turn directions to a lecture hall or guide tours at Dartmouth’s Hood Museum of Art.

As unexpected icing on the cake, the open APIs in the new AI technology are stimulating creativity by enabling the creation of cool new in-house apps that leverage self-healing capabilities.

At a time when students, faculty, and staff have never been more demanding when it comes to on-campus internet service, we’re leveraging the latest technology advances to give them what they expect and deserve.

For more information on how to architect an AI-driven campus network, watch this webinar where the Dartmouth team chats with experts from Mist.

Related Resources