WLAN Configuration For Internet Of Things
In my last post, Designing a WLAN for the Internet of Things? Think Capacity, I summarized the importance of designing the WLAN for capacity. In this article, I discuss the importance of configuration options, and importantly, what some of those options are and how they can help us create high performance networks to support the demands of IoT.
The goal of this blog post is to act as a reference for the products, features, and components that can help optimize the provisioning and configuration of your WLAN deployment. This would be a structure to make a cohesive deployment flow and provide a set of best practices. However, that said, it is by no means the exact order you must follow. Nor should every configuration option be selected in all circumstances regardless of the requirement. Yes, a feature that a vendor offers is an option, but it is not always enabled by default and you have a tick box (for example) to enable it! So you must perform simple steps first, and then reiterate as you gain more knowledge of your environment then adapt.
So should I go and enable a load of different features? Although this can be a tempting thing to do, I would not recommend it. Blindly enabling features can actually have the negative effect to the WLAN or connecting clients if used in the incorrect way. So this means we should understand the options a vendor has available and relate these to features that will aid our WLAN performance or improves bandwidth. This means we are enabling features based on our design or from something observed during a site survey or visit.
Now let’s look at the different options that are available and how these can aid our WLAN network when designing for IoT (Internet of Things).
During the site survey or capacity planning you will have identified the type of client radio that you will be supporting on the WLAN. This is key to deciding what data rates you need to leave enabled or ones you can disable. Disabling the lower data rates (i.e. 1, 2, 5.5, 11 Mbps) improves the overall capacity and performance of the WLAN, it also can remove the need to support legacy protection mechanisms (i.e. RTS/CTS).
Firewall, Rate Limit and Queuing
It is best practice in networks supporting BYOD and IoT to apply a maximum throughput limit to ensure that client devices do not hog all of the available bandwidth (for example those using BitTorrent or content rich streaming sites). Most vendors provide both firewall features alongside rate limit and queuing features. This enables you to block network services or applications you don’t want to support.
Also you can decide to rate limit traffic for a specific user, some vendors also take it a step further and allow you to create your own custom queuing policies and provide rate limits to only certain types of traffic (i.e. rate limit only YouTube traffic). It is always recommended to enable this type of feature to safeguard bandwidth.
AP Transmit power
It is best practice to ensure the AP transmit power level aligns with the power levels established during a site survey, doing so will ensure you minimize co-channel interference whist proving adequate coverage. However, many manufactures (like Aerohive) provide mechanisms where APs can automatically select the channel and power settings, this feature is often referred to as radio resource management (RRM)
Auto channel selection
As stated in the previous section we can set optimal channel (and power) for operation using some kind of RRM feature. Most RRM features have an algorithm with will evaluate metrics such as channel utilization, error rates, Wi-Fi neighbors, channel utilization to determine the optimal channel/power setting, but can also react to changing RF conditions.
Aerohive has a feature called Automatic Channel Selection Protocol, more information available here.
This is an appealing setting for people to consider adjusting as the peak data rate and throughput for single clients (that are capable of this operation) will be increased. However, we have another consideration to take around spectral efficiency, so increasing the channel width can have a detrimental effect and reduce overall network capacity.
So in a home environment, it’s ok to use wider channels as you will likely we supporting fewer devices. However, in an enterprise environment supporting company devices, BYOD, and IoT, wider channels may not be a good idea!
The preamble is used in Wi-Fi to communicate to the receiver that data is being sent. It will allow the receiver to obtain the wireless signal and coordinate itself with the transmitter. It contains additional information for identifying the modulation scheme, transmission rate, and length of time to transmit an entire data frame. The options are usually long, short, or auto.
The guard interval is described as a space between symbols being transmitted. The idea of this space is to eradicate inter-symbol interference (ISI), which happens when some type of reflection from one symbol interferes with another.
There are two options for the guard interval - long or short. Adding time (long interval) allows for reflections to settle before the next symbol is transmitted. Reducing time (short interval) increases throughput (by around 10%), but can increase the chances of ISI. The environment the WLAN is being installed and the devices connecting will dictate what setting is best used.
Dynamic Frequency Selection
Wi-Fi uses both the 2.4GHz and 5GHz spectrum. Some radar systems also use the 5GHz spectrum that these wireless devices might be operating in. So what is required when you want to utilize the full 5GHz spectrum? We enable something called Dynamic Frequency Selection (DFS), which is able to detect the radar activity and switch the channel of the AP automatically to avoid interfering with any radar systems operation.
If an AP detects such an issue it will automatically switch the AP to a different channel, and that first channel is blacklisted and unusable for 30 minutes. So if your Wi-Fi network is near a radar system, it would be best to blacklist the channels in 5GHz to avoid this issue.
As the 5GHz band offers more capacity than 2.4GHz, it makes sense to have this setting enabled as the impact of disabling DFS channels will have an effect on the capacity of the network and channel reuse plan.
Aggregate MAC Protocol Data Units
The Aggregate MAC Protocol Data Units (A-MPDU) is a technique of combining multiple frames before transmission. Also referred to as frame aggregation, the benefit A-MPDU offers is a reduction in protocol overhead due to medium, synchronization and 802.11 protocol headers. As a result, your WLAN will be more effective in operation and provide a higher overall throughput and capacity.
The frame burst feature permits a device to transmit a sequence of frames in a rapid succession. A frame burst consists of multiple, individual frames separated by a short interframe space (SIFS), which is a short amount of time normally reserved for separating a data frame from its immediate control frames.
If you decide to enable frame burst, the access point just uses a SIFS to separate data frames, thereby increasing data transmission efficiency, which results in higher throughput.
Transmit beamforming is a process of modifying the shape of its transmissions so that the receiving client obtains the best possible signal. Having a stronger signal at the receiver end will improve the client SNR, which directly affects the supported data rate the client device can receiver. The net result is higher throughput. Matthew Gast wrote a nice blog piece on beamforming (available here).
As we have already learned through this series, the 5 GHz frequency offers much more spectrum. However, most new clients that support 802.11ac most are still dual band, and also have a 2.4 GHz radio available. Therefore, it is often advantageous to encourage dual-band clients to connect to a 5 GHz radio where superior capacity exists.
In a network that supports IoT, and where APs are likely to be deployed in close proximity, the band-steering feature enables greater use of the 5 GHz band, resulting in greater capacity and a reduction in co-channel interference.
Neighboring APs will perform band steering by suppressing responses to probe and association requests on their 2.4-GHz radios to clients that are also probing in the 5 GHz band.
Most implementations provide options on how this feature, here is an example of how this helped an Aerohive customer in the past.
It is very possible in a busy network for an AP to become overloaded by client devices. Therefore, it is important to have load-balancing features that allow APs to manage association requests and clients.
Take for example a university lecture hall where you may have up to 150 users, each with multiple devices. In this type of service area, it is likely that you have more than one AP serving these clients. However, if you have an AP near the entrance to the lecture hall, it is likely that many of the clients will connect to this AP and remain connected. It easy to see how this AP could become very congested!
Load balancing enables a distribution of clients across all APs, thus maximizing the available spectrum within that service area. In this kind of circumstance this setting is very useful in providing spectral efficiency.
Dynamic Airtime Scheduling
The way in which WLAN medium contention procedures work, each station gets an equal opportunity of transmission for all frames of the same QoS class. This is because every frame of the same QoS class has a statistically equivalent probability of transmission.
However, what this does not take into account is that Wi-Fi stations transmit at different data rates, and the net effect is devices then use different amounts of airtime. For example, slower clients (802.11b) will use significantly more airtime to transmit the same amount of data as much faster clients (802.11ac).
Therefore, standard medium-contention procedures result in the slow clients hogging all the airtime and decrease the capacity for the faster clients. Many vendors therefore created an airtime-based QoS system to overcome the challenges and make airtime access more equitable.
Next in the series we will look into the solution-design phase. This is where we couple all of the three design stages into one overriding document. The development of the solution design and architecture begins with a design process, the results of which become the functional specification. Stay tuned …..
All posts in this series: