AI/ML, Automation are few terms that every vendor is talking about in the tech industry. Each vendor has their own flavor of AI/ML engines with different feature sets. Regardless of their path end goal is “better end user experience”.
Managing and operating a wireless network for thousands of users across different sites can be quite cumbersome. Mission critical wireless networks can be very challenging and demanding to operate and maintain. Organizations can not constantly send field engineers out to handle every single wireless ticket that comes in. Helpdesk or first level support agents normally do not have the skills to perform deeper Wi-Fi troubleshooting.
This brings up few questions:
1- How do you measure user experience?
2- How do you proactively find issues and fix them?
Access Point Vendors:
Sometimes, there are access point vendors that have built in intelligence and additional radios that provide insights and actionable data. This highly depends on the vendor and their AI/ML engine.
Second option is wireless sensors. We can install these sensors as an overlay to test and collect data, and provide Key Performance Indicators and Actionable Insights. I tested and used one of these sensors in the lab and production environment recently.
This sensor looks like an average access point. It is PoE powered and can easily use any PoE capable switch.
Wyebot has a total of 3 radios. The 3rd is a multi purpose radio, its used for running network tests or if theres no POE available, it can be used to connect to the internet.
I plugged it into one of the switch ports and it was up and running in few minutes. There are multiple ways to configure the Radio 1 and Radio 2 interfaces on the sensor. I left all the defaults for my testing. Screen shots below show that the Band, Channel, Channel Width are all configurable options.
I’ll go through different menu options that are available:
Device Dashboard gives you an overall view of your wireless network. Detected Access Points, Clients, Analytics, all the configured tests will show up here, these can be further investigated by clicking on the the test.
I always like to look at Wi-Fi Clients section because it gives great information about all the clients these sensors detect. This screen also allows you to look and search for the clients as well.
If there is a need to investigate any client related issues, that can be accomplished from this screen, by clicking on the client.
My SSID menu shows all the SSID’s sensor detects. You have the ability to manually add the SSID(s) or just pick one from the left and add it under the SSID Profiles.
Different authentication methods are supported:
Device Forensics, allows you to search for a client by its MAC and name (note: you can assign custom names to detected client devices)
Network Tests is where multiple tests can be setup so that the environment can be monitored. NOTE: Dual, wired and wireless tests are supported
I really like the DNS/DHCP and App tests, but I think “LinkDoctor” test is a very comprehensive test. It will test the full connectivity to ensure there are no issues in any phase. From client association , DNS, DHCP to internet access.
This information can be very helpful when trouble shooting issues; for instance if see a long Duration under Wireless Connection or long DHCP and DNS responses you are able to further investigate those issues proactively. If the DHCP requests are failing, this test will display the failure and can be further investigated.
Now my personal favorite under Advanced, Trace Download. This actually allows me to download frame captures from the local buffer. From what I can tell, I can do 24 hour time period but the USB port on the sensor allows me to hook up a larger capacity drive for additional pcap storage.
I have used Wyebot extensively for surveys and remote troubleshooting; I’d like to mention two specific issues that I had to troubleshoot remotely using Wyebot:
Second issue was also related to multiple clients having poor performance. Using a combination of frame captures from the Wyebot sensor and the information reported in the dashboard I was able to identify the issues remotely.
One important thing to note here is that all this troubleshooting was done remotely. It saved time and money; I did not have to fly out to these sites and spend countless hours trying to replicate the issue and capture data manually. All this was done by Wyebot sensors. Not only this saved my time and money but added to client satisfaction. Leveraging the benefits of AI/ML and sensors can drastically improve how well you are able to support and maintain your WLAN environment.