KodaCloud CEO Explains How AI Improves Wi-Fi

The Adaptive & Intent-Based Networking Expo is less than three months away. So now would be great time to register for this new TMC event, which was designed to help business transformation, and business application and networking leaders and team members, learn more about how artificial intelligence- and machine learning-based solutions can help them automate network operations, address the data deluge, scale, secure their networks and applications, and become more agile and innovative.

Wi-Fi networks are one area in which AI allows for marked improvements. Bernard Herscovici, founder and CEO of KodaCloud, will explain how during The Adaptive & Intent-Based Networking Expo panel “How to Use AI to Take Wi-Fi to the Next Level”.

Prior to founding KodaCloud in 2014, Bernard was vice president of Wi-Fi at Ericsson (News - Alert), and founder and CEO of BelAir Networks, a company that was sold to Ericsson in 2012. Prior to BelAir, Bernard was the president of Breezecom Inc. and vice president at Newbridge Networks.

Bernard has co-authored several patents and has an MBA from the University of Ottawa and Master of Science in Wireless Communications from the University of Toronto (Canada).

Why did you establish KodaCloud?

Bernard: The existing Wi-Fi experience for business is broken. So we started down the path of reinventing how Wi-Fi was managed and operated with quality of user experience at the core of everything we do. From this emerged the vision for AI-driven cloud service Wi-Fi.

What does KodaCloud sell?

Bernard: We offer a Wi-Fi on an as-a-service model. Customers may use KodaCloud-supplied access points or use Aruba, Cisco (News - Alert), or Ruckus access points with the new eNOC service which KodaCloud will manage using its cloud-based advanced AI solution.

Who needs this kind of solution and why?

Bernard: Any business that wants to have more reliable and significantly better performing Wi-Fi. This is particularly important if the Wi-Fi network is used for mission-critical reasons; for example, oil and gas fields, hospitals, manufacturing processes, retail (inventory tracking), etc.

You’ll be speaking at The Adaptive & Intent-Based Networking Expo early next year in Fort Lauderdale. This new event will focus on how artificial intelligence, automation, and machine learning can allow for better performing, more agile, secure, and cost efficient networks, applications, and businesses. How does KodaCloud fit into that discussion?

Bernard: This is precisely what KodaCloud does. It is really the first step in automating IT by removing the need for specialized Wi-Fi IT support personnel, resulting in better costs and better results as AI (and the underlying ML) can look at all devices connected to all access points simultaneously in real time and make decisions and implement changes at machine speed rather than human speed.

KodaCloud in May expanded its portfolio with the addition of its eNOC service. What is eNOC, and why is it important?

Bernard: eNOC allows customers to use third-party access points from Aruba, Cisco, and Ruckus to access the KodaCloud AI service. eNOC focuses on the overall QoE for all devices (end users and IoT devices) connected to the Wi-Fi network. The service uses the concept of device profiles, which allow the IT department to set thresholds for different types of end user or IoT device (for example a printer does not require the same bandwidth as an HD video camera) and then have the KodaCloud AI manage to that profile.

If the KodaCloud service cannot resolve an issue, it will issue an alert to IT with a clear actionable root cause and solution. The service also stores all interactions with all devices in a historical database so IT can easily go back and see the performance of all access points and all devices connected to those access points which negates the typical problem of having to re-create an issue in order to fix it.

Does eNOC leverage AI?

Bernard: Yes, this is a critical component of the service as AI combined with a big data environment allows all devices connected to all access points to be monitored in real time and enable the AI to analyze any anomalies and make changes or recommendations at machine speed, thereby often fixing issues before they become user affecting.

Is there a difference in your mind between AI and machine learning?

Bernard: Yes. The AI solution can be considered as an infinite number of tier 2 Wi-Fi experts making changes based on their expert knowledge. The ML portion is responsible for learning, for example, that thresholds should be adjusted for certain devices as the newer threshold will result in better overall connectivity or performance.

How would you describe the level of sophistication of AI-based solutions today?

Bernard: Relatively early but sophisticated enough to make real, significant improvements to networks as compared to using human IT.

What do you hope and expect to become possible with AI?

Bernard: Ultimately we believe all networking (LAN, WAN) will be delivered as a service with AI running the service, freeing IT to focus on differentiated technology for their business rather than generic networking.