A couple of months ago I was having dinner with a fairly well-known Silicon Valley executive who predicted that success for an IT vendor is based on two things: having lots of data and a robust artificial intelligence (AI) engine to discover new insights.
If that is true, then Mist Systems seems to be in a strong position, as the company’s solutions were designed to use AI to solve some of the bigger challenges in Wi-Fi today.
This week the wireless network company announced several new access points, as well as use cases, for its solution. Specifics are as follows:
Introduction of client service-level expectations (SLE)
In telecommunications, the concept of a service-level agreement (SLA) is a threshold that service providers are contracted to meet. The SLE from Mist is similar, although more proactive than a carrier’s SLA. With Mist, administrators can use data to set, monitor and enforce things that impact performance pre and post connection. Examples of this are time to connect, failed connection attempts, roaming, coverage, capacity and AP uptime. The SLEs can be monitored in real time and watched over time to provide up-to-the minute insight as to the health of Wi-Fi.
The SLEs can be applied at multiple levels, including at a site, access point or even down to the individual client level, giving IT the ability to track and optimize the performance of every individual on the network. Mist uses microservices to achieve this level of granularity, enabling its control plane to track the state of each client and then process the massive amount of data using its machine learning algorithms.
Improved anomaly detection
The SLEs enable network managers to understand what normal operations looks like. Mist aggregates the data and analyzes it to look for anomalies that could indicate something that negatively impacts experience before the users start calling. My research shows that 75 percent of performance problems are actually caught by the end user and not the IT department, which is why there is so much frustration from workers towards company help desks.
Gross deviations are easy to spot with the naked eye, but ones that span APs or involve multiple infrastructure components are hard to find, which is where AI has its value. Mist is able to proactively identify deviations before they impact user experience, putting IT in the drivers seat. Mist looks for anomalies across devices, operating systems, APs and applications. Mist uses this primarily for performance monitoring, but anomaly detection can also be used to spot security breaches. I expect to see Mist add this in the future.
Personal WLANs using virtual network segmentation
Mist is leveraging its current WxLAN policy engine to enable the secure onboarding and segmentation of users and Internet of Things (IoT) endpoints. Currently, if a company wants to securely share resources that don’t support 802.1x authentication, individual pre-shared keys must be used for each group. Mist is introducing something called personal WLANs that allow for onboarding and segmentation of multiple users and devices on a common SSID, using personalized pre-shared keys across the network.
With a personal WLAN, a user could create his own private key, apply it to a device and then share it with a group for access to it. This feature lets individual users create their own personal, virtual networks, making them ideal for environments with lots of users that need to share devices, such as hotels, schools and malls.
New access points
Mist added the following to its family of APs:
- AP21—High-performance converged Wi-Fi and BLE AP that supports 802.11ac, 2×2 Gigabit Wi-Fi for dual-band services and a directional 16 element Bluetooth low energy (BLE) antenna array that supports the company’s location services. The AP21 is available today.
- AP61—Outdoor 802.11ac wave 2 4×4 Gigabit WiFi with directional 16 element BLE antenna array. The AP61 will be available in Q3 of 2017.
Since its launch, Mist has talked the talk of AI and Wi-Fi and built its products with the vision of a world where everything is connected and that data driven from these connections can create new insights through the use of AI. These new features are a great example of new innovations that can be brought the wireless network so it can handle a world where billons more devices will be connected over the next several years.