Mist wireless analytics service captures NIA

Mist makes plans to diversify its innovative wireless analytics service, as the company grows its customer base to more than 200 users.

This month's Network Innovation Award goes to Mist Systems, which in 2016 rolled out a wireless analytics service that blends artificial intelligence, patented virtual Bluetooth Low Energy technology and cloud-based machine learning to provide enterprises with the information they need to monitor the performance of their wireless LANs.

SearchNetworking spoke with Mist CEO Sujai Hajela to see how Mist has fared in its first year of operation, and to find out what services and capabilities the company plans to roll out in the years to come.

This interview has been edited for length and clarity.

What was the market and technology niche you were trying to fill with Mist when you launched the service in 2016?

Sujai Hajela: You have to look at the evolution of the Wi-Fi market. In the early 2000s, it was really about portability. When we began looking at developing Mist [earlier this decade], frankly, there had been hardly any innovation in the space for more than a decade. At the same time, our mobile needs had continued to grow. It wasn't about mobile devices just getting connected. It was more about [mobile] experiences. With Mist, fundamentally, we're ushering in a new wireless network that personalizes the user experience and enables smart mobile devices to provide these personalized experiences.

What does that mean?

Sujai Hajela, CEO, MistSujai Hajela

Hajela: If I'm a passenger at an airport, if I'm a guest at a hotel, if there is an unfortunate incident and I am a patient at a hospital, our mobile devices become the digital interface to the world, and what we are doing is making sure those mobile devices give me the experiences I need and rely on.

How does Mist wireless analytics accomplish that? Is that where the machine learning and automated analytics capabilities play a role?

Hajela: Those are critical. If you look at a personalized mobile experience, what does it entail? When I mentioned earlier about being a guest at a hotel, that would mean being able to engage with that enterprise. It would also mean allowing the hotel's IT team to be proactive in managing what's happening with that experience.

Today, in most cases, network IT has no idea about what's happening with users' mobile connections. So, on what basis can an experience be provided if IT doesn't even know what that experience is? The only way you can begin looking at network performance, to offer a personalized experience, is to put artificial intelligence and machine learning into play.

Without that, it's impossible for [an organization] to manually track the zetabytes of data hitting the network. We are taking this huge amount of information, which is coming onto the mobile network, and then processing this information using artificial intelligence to get insights from the network to deliver that personalized experience. 

Mist has been in operation for just a little more than a year. How many customers does Mist have?

Hajela: We have 200-plus customers, 100 partners and have seven of the top Fortune 20 companies using us to provide either wireless assurance or location tracking through our virtual Bluetooth patented technology we released. In calendar Q3, which hasn't ended yet, we are experiencing close to 500% quarter-over-quarter growth. In general, we are way ahead of the plan we had committed to the board for this year.
Are the use cases what you expected them to be, or are you are finding that some customers are using this wireless analytics service in ways that perhaps you hadn't anticipated?

Hajela: I will say we hadn't anticipated some of the use cases. When we went right in, it was all about, 'How can we use artificial intelligence to provide insights for network IT?' Now, [we're seeing Mist being used] for asset tracking, as well as way-finding and proximity messaging tracking through virtual Bluetooth Low Energy beacons and Wi-Fi.

It was amazing how some of our key customers leveraged what we were providing in terms of wireless assurance and location-based services to come up with very unique use cases. One of those use cases was all about augmented reality. It was like the Pokémon Go for the enterprise.

What are some other examples that the beacon technology is being used?

Hajela: Industrial uses. As you know, especially with heavy machinery and all that, it's very difficult to carry manuals [and other necessary information]. So, what some companies are doing is using the virtual BLE technology to have the manual or catalog content exactly where the worker might be. And while he's at the machine, the worker is wearing glasses that are actually painting a picture of what he might need to do. So, we're seeing augmented reality coming into play.

Normally, amusement parks will scribble things on walls for you to read so you get excited about characters or the ride. But my kids aren't reading the walls; they're actually on their phones.
Sujai HajelaCEO, Mist

A very simple example is at amusement parks. Imagine you're waiting in line, and these waits can go 45 minutes or longer. Normally, amusement parks will scribble things on walls for you to read so you get excited about characters or the ride. But my kids aren't reading the walls; they're actually on their phones. So, what these parks are doing is experimenting with technology that says, 'Would you like to interact with this character?' You click on your device, and now you're getting into an immersive experience with that character being served through the virtual beacon.
What else are you doing to accelerate the development of the virtual BLE technology?

Hajela: We are converging Wi-Fi and BLE, which ensures that not only can that [particular mobile] experience be served to you, but network IT is clearly seeing what's going on.

Earlier this year, Mist added support for network segmentation. What are some of the areas in which you see this capability being a benefit to enterprises?

Hajela: One of the first is personal WLANs, where you can create multiple, personal networks from the same overall wireless network. Basically, you could go into a hotel, and each room would have its own network; or in a dorm, each room, again, could have its own network under the same service set identifier.

Second, we are leveraging this for security, along with artificial intelligence. Our system can see, for example, [if someone has spoofed a MAC address] and is trying to get on the network. Now, we're using location accuracy, through the convergence of Wi-Fi and BLE, to determine whether someone has access to a certain service or not.

Can you be more specific on how Wi-Fi and BLE would work in this scenario?

Hajela: Let's say you have an enterprise. Its wireless network has a green zone, which is where it's OK for everyone to have access. Then, it has an amber zone, where contractors should not be able to do certain things. And it has a red zone, where, let's say, a phone's camera should be disabled. How do you guarantee that?

We're leveraging location-based technologies through BLE to enable that experience. So, from a security perspective, not only are we using AI [artificial intelligence] to detect anomalous behavior and to detect anomalous connections, we are adding location information to enable security that is now location-based. So, it's not just segmentation based on who you are and what you're using, but segmentation also based on where you are.
Let's talk a little bit more about the end-user experience. How are enterprises using wireless analytics to measure end-user experience more specifically and proactively?

Hajela: What we are doing is monitoring your device, every second, for all the types of information coming from the device. Then, we are constructing what we call a simple state machine. The system is constructing this state machine to answer questions like, 'Are you able to connect or not? What is your coverage? What is your capacity? What is your throughput?' It then constructs millions of these state machines, and then it is watching for anomalies. If it encounters an anomaly, it checks various [parameters]. Is it a device problem? A network problem? A cloud service problem? It then breaks down the problem, and then it identifies the applications that are being used. I can then know if it's a wireless connection issue, or if my application isn't working, or if it's the WAN or the cloud.

Next Steps

augmented reality and the enterprise: What's next?

Network segmentation and its role in security

Say hello to advanced analytics

This was last published in August 2017

Dig Deeper on Wireless LAN (WLAN)