Blog
Wi-Fi roaming issues are frustrating and often relate to a few key things, such as poor network design, where there is excessive cell overlap due to high access point power levels. As a result, client devices "stick" to the distant cell. Another culprit of Wi-Fi roaming issues is RF bleeding between floors, where clients connect to an access point that has a very small coverage area, therefore, the signal degrades rapidly. Finally, there are issues with client devices themselves, because in Wi-Fi, the devices always make the decisions about roaming. The device manufacturer, model or software version must all be considered. Beyond upgrading or changing devices, some Wi-Fi roaming issues may only be fixed by the manufacturer.
7signal is currently developing a new Wi-Fi roaming test, which would move sensor connections between APs and measure roaming delays. We see this as an important new feature, which would allow us to continuously benchmark Wi-Fi roaming performance and behavior offered by network.
However, we already have a number of powerful metrics to observe Wi-Fi roaming behavior. This data is based on automated passive captures where we calculate metrics on access point/client device interactions. The best results for roaming analysis are achieved when Eye sensors are configured to run only passive tests on one SSID for the duration of the Wi-Fi roaming troubleshooting. This allows the maximum amount of samples to be collected.
Below is a list of the important data elements and what to watch for when identifying Wi-Fi roaming issues:
Retry Rates
High retry rates indicate poor RF connections. Identify specific device types or individual devices, which have issues. Also, observe uplink and downlink retry rates separately.
Data Rate/MCS Distribution
Low data rate/MCSs is another great indicator of a poor radio connection. Check for a proper radio-link balance by comparing uplink and downlink data rates. High output power from the access point only helps in the downlink direction. It does not help when the device is talking to the access point. In fact, when access point power is too high, it makes most devices think the connection is good, therefore, it does not want to roam. Individual devices and device models vary in their capabilities. So, by measuring, you can identify the devices struggling with Wi-Fi roaming issues.
Devices Connecting to Access Points in Adjacent Areas
Often terminals are supposed to operate on only one floor or area. If you catch a device “talking” with access points on different floors, then you likely have an RF leakage or propagation issue.
Number of Devices Connected to an Access Point
Observing the number of devices connecting to certain access points at certain times of the day also provides excellent insight. Staff meetings and other gatherings provide nice opportunities to observe roaming patterns.
Probe Requests and Probe Responses
The number of probe requests and responses and the balance between them also provides valuable insight into Wi-Fi roaming issues. An excessive amount of probe requests by a device may mean it is lacking the sufficient received signal levels, or is having trouble finding an access point to roam to.
Re-association Frames
Re-association frames provide direct information on roaming behavior for each device and device manufacturer. For example, you can identify the number of re-association requests and responses for different device manufacturers and individual devices. Also, re-association response codes indicate whether Wi-Fi roaming was successful or not successful.
Steps to Understanding and Resolving Wi-Fi Roaming Issues
- Benchmark the performance of your Wi-Fi network in the problem area to ensure it has the capability to handle the traffic. It is best to do this during periods of both high and low traffic. It’s important to first ensure that all access points are up and have the ability to offer continuous connectivity.
- Review your network design to identify cell overlap and bleeding between floors and areas. This may require a fresh survey with your preferred survey tool. When viewing the results, remember to adjust coverage areas based on your device's actual received signal level. Survey sticks usually have better antennas and positioning, therefore, generally show higher signal levels than handheld terminals. Adjust the access point power levels for sufficient, but not excessive, overlap. Often times, 30% overlap with -65 dBm as a cell border is a good starting point.
- Review client device settings and its physical positioning. Sometimes devices may be positioned so that their Wi-Fi antenna performance is compromised. Some devices have settings impacting Wi-Fi roaming behavior.
- Review client specific behavior from passive sensor data. First, give an alias for the devices most interesting to you. If you know the users who are experiencing the Wi-Fi roaming issues, then you can correlate expected user behavior with the sensor data provided. Compare devices and device models. Observe anomalies and outliers based on metrics.
- Make adjustments to your network and your devices. You can impact client behavior with different of settings in the network, including access point power levels, enabled data rates, SSID allocations, supported bands, supported standards, band steering and load balancing settings. Finally, you may also perform some physical changes, like moving APs or using different types of antennas, which provide more focused coverage area. Check the impact of your changes by running more tests and conferring with the data.
After making these kinds of adjustments, you will gradually hear less Wi-Fi roaming complaints from users. Wi-Fi roaming issues can be tricky to resolve and even more difficult to find without data to understand how clients interact with your network.