Mist Systems Enhances Wi-Fi Services

Marvis expanded NLP: When Mist first rolled out Marvis, its AI-driven virtual network assistant in February, it understood a minimal number of questions that drive insight into WLAN behavior and offer guidance for Wi-Fi troubleshooting. With its continuous learning capability, Marvis can now understand and answer hundreds of requests, such as, “What is wrong with Jon’s iPhone?”, “How is the access point in the conference room doing?” and “How many devices are having throughput issues?”

Anomaly detection with deep learning: The latest version of Marvis uses deep learning to identify and correlate items, events and other observations that do not conform to expected patterns and predict future events based on the sequence of past states. By leveraging advanced data-science algorithms, Mist detects and reports anomalies with almost zero false-positives. The Mist Learning WLAN is 100 percent API-driven, enabling workflows, such as help-desk tickets and notifications, to be automatically kicked off when anomaly thresholds are triggered. This helps IT staff get ahead of problems and fix them before users even know they exist.

AI-driven radio resource management: The Mist Wi-Fi Assurance Service is constantly collecting RF information, such as coverage, capacity, throughput and performance on a per-user basis, and then uses deep learning to understand the data and make automated changes based on real-time requirements. Customizable Wi-Fi service levels measure the impact of these changes so that the user experience is always optimized. For the first time, the wireless network can detect RF issues, learn from them, automatically make the changes using AI, and then measure the results as quantifiable end-user benefits.

WAN SLEs: Mist offers customizable Wi-Fi service levels, which have traditionally covered key metrics like capacity, coverage, throughout, latency, access point up-time and roaming. Now, the Mist platform also lets IT administrators set, monitor and enforce SLEs for WAN performance, allowing customers to determine if WAN latency, jitter and/or packet loss are having an adverse impact on the user experience, when these problems are occurring, if they are likely to happen in the future, and how best to address them for an optimal user experience.“This is all about helping our partners get to the root cause of an issue quicker and to be able to set up service levels where they can actually guarantee a service and measure the service that they’re delivering to their end users,” said Aaron. “This provides user visibility into the network that didn’t exist before.”