Recent Posts

Posts by Topic

see all

Natural Language Interface for Mojo Cognitive WiFi

by Murthy Jvr on Feb 20, 2018

You may have experience asking the digital assistant on your smartphone to schedule a meeting or provide weather update. What if you could have a similar assistant to ask about the overall health of the network, debug client issues or take corrective actions?  

A wireless network platform with such a capability can greatly assist in day to day management and in finding and fixing complex issues. Combining natural language processing techniques (NLP) with the intelligence of Mojo Cognitive WiFi™  that continuously monitors user experience, detects potential issues and provides solutions, has made this a reality as I will describe and demonstrate in this blog.

At the foundation of Mojo Cognitive WiFi is its cognition plane that spans across the intelligent edge access points and the Big Data and machine learning platform in the cloud. The cognition plane provides various functionality such as:

  • Granularly track each step of the client in great detail
  • Measure the quality of experience of a client through various metrics
  • Proactively detect issues in the network
  • Learn patterns of network usage
  • Track application usage and measure quality of service
  • Identify root cause of issues and provide recommendations or solutions

By integrating NLP into the Cognitive WiFi platform, network administrators can now have a natural language interface that serves as an assistant in carrying out routine tasks or in gaining insights into complex issues. The natural language interface can understand the intent and associated parameters of the actions that IT administrator wants to perform through natural conversation and respond by providing relevant information, insights and root causes. Example questions that a user can ask are as follows:

  • How is the wireless network at Corporate?
  • Troubleshoot the issue with Robert’s Macbook?
  • Troubleshoot the low data rate issue near cafeteria.
  • Let John connect to the guest network for 2 hours.
  • Are there any voice call quality issues today?
  • Are you seeing any coverage issues on second floor?

The intent and parameters extracted from the high level questions are used to carry out the desired action, or to mine and present the most relevant problems and solutions, thereby quickly fulfilling the requests. This enables the administrators to get to the root cause of issues and the final solutions on a fast track and maintain a great WiFi experience for their users.

So what does it all mean for WiFi administrator? Is this just some fancy feature or does it have real benefits? I would argue it is the latter. For one, with the natural language interface, the admin does not have to search or spend time on getting from his intent to a place in the management console where relevant information and configuration options are presented. That is the navigational aspect. On top of it, with the cognition plane doing much of the analysis, the administrator can get directly from intent to inference based and actionable information. So I expect it to result into massive productivity gains.

Check out this video for the demonstration of this capability.


Topics: cloud, WLAN, WLAN networks, Wireless security, featured