Radio Resource Management (RRM)
SUMMARY Mist RRM allows the APs to participate in the fine-tuning of the network through reinforcement learning. Using RRM, the APs in a site can adapt automatically to interference and capacity issues to ensure exceptional user experience.
When talking about radio resource management (RRM), most wireless AP vendors focus solely on channel reuse. This single-factor approach fails to consider other elements that affect the user experience or the dynamically changing daytime wireless environment when clients are:
-
Making new attachments and authorizations
-
Roaming around the environment
-
Transmitting and receiving data
-
Using some percentage of available capacity
-
Trying to overcome multiple sources of wireless interference
In addition, the traditional approaches to RRM contain no means to determine if changes to the spectrum have had any effect.
Juniper Mist RRM has been designed to prioritize the user experience. Mist RRM utilizes a reinforcement learning-based feedback model which takes into account key factors from the Wireless Capacity SLE including client count, client usage, and interference. Mist RRM can automatically adjust AP power or change wireless channels when the capacity SLE is not met. After the change, Mist continues to monitor the capacity SLE to determine whether these changes in channel or power produced measurable improvements, thereby providing a better user experience. This auto-tuning process is continuous.
In other words, Mist RRM makes the wireless network better over time.
Radio frequency environments are inherently complex and therefore challenging to control and optimize for the efficient transmission of data. Since the inception of radio frequency, or RF, radio resource management, also known as RRM, has been a long-standing technique used to optimize the RF radio waves that transmit network traffic in wireless LANs. However, multiple interference sources like walls, buildings, and people combined with the air serving as transmission medium make RRM a challenging technique to master.
Traditionally, site surveys have been used to determine the optimal placement of Wi-Fi access points and settings for transmit power, channels, and bandwidth. However, these manual approaches can't account for the dynamic nature of the environment when the wireless network is in use, with people and devices entering or leaving and moving about. Additionally, this challenge is compounded with random RF interferences from sources like microwave ovens, radios, and aircraft radar, to name a few.
But what if the wireless network itself could perform RRM on its own? What if it could detect and respond to both interference sources, as well as the movement of people and devices, and adjust the radio settings in real time to provide the best possible wireless service? That's exactly what Juniper has done with the AI-driven MIST wireless solution, using advanced machine learning techniques. Specifically, MIST uses reinforcement learning to perform RRM. In a nutshell, a reinforcement learning machine, or agent, learns through an iterative trial and error process in an effort to achieve the correct result.
It's rewarded for actions that lead to the correct result, while receiving penalties for actions leading to an incorrect result. The machine learns by favoring actions that result in rewards. With MIST wireless, the reinforcement learning machine's value function is based on three main factors that lead to a good user experience.
Coverage, capacity, and connectivity. A value function can be thought of as an expected return based on the actions taken. The machine can execute five different actions to optimize the value function.
These are adjusting the band setting between the two wireless bands of 2.4 GHz and 5 GHz, increasing or decreasing the transmit power of the AP's radios, switching to a different channel within the band, adjusting a channel's bandwidth, and switching the BSS color, which is a new knob available to 11 AX access points. RRM will select actions with maximum future rewards for a site. Future rewards are evaluated by a value function.
The various actions taken by the learning machine, such as the increase of transmit power or switching the band from 2.4 GHz to 5 GHz, together represent a policy, which is a map the machine builds based on multiple trial and error cycles as it collects rewards, modeling actions that maximize the value function. Again, keep in mind that the value function represents good wireless user experience. As time goes on, even if random changes occur in the environment, the machine learns as it strives to maximize the value function.
The benefits of using reinforcement learning are obvious. A MIST wireless network customizes the RRM policy per site, creating a unique wireless coverage environment akin to a well-tailored suit. While large organizations with multiple sites replicate their many locations as copy exact, these sites will naturally experience variances despite best efforts.
Reinforcement learning easily fixes this, delivering real-time, actively adjusting, custom wireless environments. We hope this episode helped to uncover some of the magic and mystery behind our AI-driven network solutions.
How Juniper Mist RRM Works
Juniper Mist RRM measures and calculates capacity, usage, and interference factors all day, every day. RRM uses these calculations and measurements as references to the users' network experience, also known as user minutes. RRM stores up to 30 days of this data which creates a long-term trend baseline. Using regularly-scheduled (nightly) and manual corrections, RRM can adjust to shortcomings or leverage enhancement opportunities in the wireless environment by:
-
Using automatic channel switching (ACS) to respond to overcrowded or interference-prone channels
-
Using automatic power adjustments to increase or decrease the AP power output (based on client experience)
-
Using auto-cancelation to disable the 2.4 GHz radio on certain APs in the network
-
Using auto-conversion to convert dual-band capable radios from 2.4 GHz operation to 5 GHz operation
RRM operates two tiers of optimization: global optimization and local, event-driven optimization.
-
Global RRM—uses long-term cloud-based trend data from the Wireless Capacity SLE to determine if a change will be beneficial. Even though global optimization gathers and analyzes data continuously, it only takes action once per day and only if the action will have a positive effect based on learned data. Changes made by global optimization are shown as a green triangle called Periodic Optimization on the chart at the top of the page: Monitor > Service Levels when you click the Wireless button. The changes from Global RRM can also be seen on the Site > Radio Management page in the Radio Events section and are labeled as Scheduled Site RRM.
-
Local (event-driven) RRM—uses the Wireless Capacity SLE. Whenever there is a deviation below the capacity SLE that also impacts users, local RRM will attempt to make a change by adjusting the power for an AP or changing the channel that the AP is broadcasting on. Local RRM makes use of the dedicated scanning radio in every Mist AP that scans all the channels all the time. It uses the scan data from each AP to build a channel score and a site score. When the capacity SLE triggers a channel change, RRM refers to the channel score for the AP to determine which channel to assign. For example, if you are seeing an issue that you suspect to be related to WiFi interference, you might wonder why two APs are both on the same channel. If you look at the capacity SLE, and see no user impact, it means that RRM is deferring the change to global RRM which will make the channel change at night. If there was negative user impact, local RRM would make the change immediately.
Table 1: Comparing Global and Local RRM Global RRM
Local RRM Is scheduled to run automatically every night, per site.
Reacts to local events that impact users.
Can be triggered manually per band with the
button on the Site Radio Management page The button text changes with band selection.
Cloud independent
Uses reinforcement learning
Runs as needed
Creates and maintains a data set that spans multiple days which it uses as historical reference.
Reacts to the following event types:
- Auto channel selection (ACS)
- Auto-triggered ACS
- WiFi interference
- non-Wi-Fi interference
- Neighbor AP down
- Neighbor AP recovered
- Radar detected
- Post radar
Neither tier of Mist RRM makes changes for the sake of making changes. If the Capacity SLE for a particular site is 90% or above, there's not much to be gained by making changes, so RRM doesn't make changes. Additionally, if a change is warranted but RRM can't make a positive change, there might be something in the environment that needs further investigation.
When Mist RRM changes channels, it does so based not only on the current environment, but on historical knowledge. Even if the current environment makes the use of certain channels look good, Mist remembers if it has seen co-channel interference, or other problems, on that channel. If so, RRM will deprioritize that channel. In essence, Mist RRM offers a regularly-updated, prioritized channel list to clients based on what it knows about the network environment at your sites. Automatic channel prioritization helps make Mist RRM less disruptive than any other RRM solution out there.
Auto Cancelation and Auto Conversion
There are two additional RRM-related features you should know about: Table 2.
Auto Cancelation |
Auto Conversion |
---|---|
Automatically disables 2.4 GHz radios. |
Automatically converts dual-band capable radios to 5 GHz operation |
Reduces co-channel interference in the 2.4 GHz band by reducing the number of broadcasting radios. |
Reduces co-channel interference in 2.4 GHz spectrum by reducing the number of broadcasting radios. |
Improves performance on the 2.4 GHz band. |
Improves performance on the 2.4 GHz band. |
Turns off 2.4 GHz radios only if the removal of that radio will not cause neighboring APs to increase transmit power to compensate. |
Converts 2.4 GHz radios only if the removal of that radio from the 2.4 GHz network will not cause neighboring APs to increase transmit power to compensate. |
Typical cancelation rate for 2.4 GHz radios is roughly 40%. Auto cancelation never removes more than 50% of 2.4 GHz radios in a given site. |
Typical conversion rate for 2.4 GHz radios is roughly 40%. Auto conversion never removes more than 50% of 2.4 GHz radios in a given site. |
Supported on all Juniper Mist APs | Supported only on AP43, AP45, and AP63 models |
Increases coverage in the 5 GHz band with the addition of another broadcasting radio |
You might want to consider auto cancelation or auto conversion in primarily 5 GHz networks where the important devices are managed and their roaming profiles are well known. In schools or other environments where you don't care about the guest network or the variety of client devices that might show up, these features can be very beneficial.
On the other hand, you might want to disable these features in less densely covered environments where a lot of mission critical devices run only on 2.4 GHz.
Dual 5 GHz Operation
When the AP43, AP45, or AP63 are operating in Dual 5 GHz mode, the radios split the 5 GHz band and are locked to a specific range of channels. See Table 3.
Wireless Mode |
Dual Band Radio (2.4 GHz) |
Dual Band Radio (5 GHz) |
5 GHz Radio |
---|---|---|---|
Dual Band Mode |
All 2.4 GHz channels |
N/A |
All 5 GHz channels |
Dual 5 GHz Mode |
N/A |
Channels 100-165 |
Channels 36-64 |
We recommend setting the 5 GHz channel width to 20 MHz when using auto-conversion or dual 5 GHz. Using the 20 MHz width helps maximize the number of 5 GHz radios in use while minimizing co-channel interference.
If you want to use operate Dual 5 GHz radios in 5 GHz mode, configure Dual Band Settings to 5 GHz and set the 2.4 GHz Settings to enabled.
![](../../images/rrm-5ghz-dual-band.png)