WiFi 6 has finally arrived in consumers' homes. If the industry forecasts hold up, this technology will be one of the leading trends of the decade.
WiFi 6 has finally arrived in consumers’ homes. If the industry forecasts hold up, this technology will be one of the leading trends of the decade. Just this year, the Wi-Fi Alliance anticipates nearly 2 billion WiFi 6 device shipments to consumers, commercial entities, and government agencies. There’s a lot of buzz about WiFi 6 transforming the consumer experience and connected environments. What’s hidden in all the excitement is one caveat: WiFi 6 is only as good as the system controlling it.
For all its breakthroughs, the technology’s new capabilities will not perform to their full potential without intelligent management and optimization of the home WiFi network. To truly improve the customer experience in the congested home-network environment, WiFi 6 requires additional tools, such as network optimization and controls.
Here’s a breakdown of some of the top WiFi 6 capabilities, the problems we can expect to arise, and how they can be solved.
160 MHz channel bandwidth
Most WiFi 6 devices are expected to support 160 MHz channels, whereas WiFi 5 had the capability but few WiFi 5 devices went beyond 80 MHz. In theory, since doubling the channel width doubles throughput, customers would experience downloads that are twice as fast, double resolution, etc.
The problem is that in most countries, the US included, only two independent 160 MHz channels are currently available, and devices that use smaller bandwidth also share these channels. This will become a huge problem for dense environments, where many competing devices or overlapping channels used by different subscribers will create significant interference.
Mitigating this issue requires specific network configuration. First, the interference on the 80 MHz sub-channels of the 160 MHz transmissions needs to be sensed and predicted, so that channel allocation and bandwidth selection can be done intelligently. In addition, it’s important to consider the complete interference picture in an environment so that channel allocation can be optimized across an apartment complex (or neighboring homes) and allocated where it’s needed the most.
This configuration can’t be achieved by locally managed networks and must be handled centrally through the cloud. In an apartment complex, for example, a cloud platform would analyze factors such as loads and device types, as well as take into account the history and current set of clients and loads that are present in the network at each access point (AP). Based on this data and analysis, the platform would then assign channel bandwidth to each AP across the entire building.
To minimize the conflicts between neighboring units, the frequency channels can also be “tiled.” And if channel reuse can’t be avoided, the optimization platform would analyze historical activity to select apartments that can share frequency channels with the least issues.
This marquee WiFi 6 feature improves efficiency and capacity by subdividing the channel into smaller frequency allocations (resource units) that are transmitted from one AP in parallel, so that a single transmission can communicate with a large number of devices. This eliminates 802.11 packet overhead and wasted time, particularly for transmissions involving IoT and other devices with low data rates.
The problem is that OFDMA only improves efficiency when a significant number of IoT devices share an AP. This is not always the case in today’s smart homes, where the network has evolved to multi-AP topologies involving mesh networks, repeaters, or multiple gateways.
Each AP would not have enough devices to group into an OFDMA transmission if clients were simply allowed to connect to the nearest AP. Yet forcing clients to connect to a distant AP will cause dropped data rates. That means OFDMA operation will need rigorous optimization and client steering.
The reasoning behind OFDMA-aware client steering is complex. It requires a centralized, intelligent network controller to know what APs and clients have WiFi 6 capabilities, to factor in historical observation data, and to make forward predictions of each client’s data needs.
The controller would then need to make optimized choices about which AP each device should connect to based on capabilities, traffic load, signal strength, and the device’s data use. And finally, the controller would need the ability to steer each device — using a steering mechanism specific to that device type—and hold it on the correct AP, even if it’s not the closest AP.
Target wake time
Target wake time (TWT) improves battery life for devices that only transmit occasionally or at a low-duty cycle. The AP reserves a time window for each client to wake up, quickly communicate, and return to sleep, while keeping the reserved air time clear of other communications.
In the home network, several APs often operate on the same frequency channel, which means multiple APs may try to schedule the same TWT for different devices. To avoid overlap, a cloud-based, central scheduler will need to coordinate TWT periods for the co-channel APs. To create an optimized TWT arrangement at each AP across the home network, the scheduler needs to know transmission requirements of all the TWT clients, the shared channels among APs, which clients connect to which AP, and the signal strength between the APs and clients.
In an apartment complex, the optimizer could additionally look across all the overlapping networks that it controls and plan TWT assignments and groups across the entire building. The controller would factor in data about which apartments interfere with one another, and the client capabilities and load requirements in each apartment.
6 GHz frequency
Both the low- and the high-power uses of the 6 GHz spectrum will present issues. Since the low-power transmission (18dBm/63mW) can’t transmit as far, or at as high a data rate, more complicated, multi-AP configurations will be required for an optimization system to select the appropriate network topologies, frequency assignments, and client-steering options. On the other hand, to use the automated frequency control (AFC) system (30dBm/1 Watt allowed), the transmission will need to avoid any frequency channel used by nearby microwave systems. This involves communication with a smart controller that can look up the FCC database, factor in the geo data, calculate interference levels, then deliver instructions back to the AP.
For either mode, the controller must take into account factors such as client types, loads, and capabilities before allocating the network’s radio resources. Depending on the capabilities of the clients in the network, it is not always optimal to put one of the AP’s radios in the 6 GHz band. For example, using a 6 GHz channel for the backhaul connection may help the backhaul, but it may take away the high performance radio in the AP from the 5 GHz band. High-performance clients that do not have a 6 GHz capability may therefore connect at lower speeds, actually degrading the experience in the home.
The home network will continue to grow more congested and complicated. The average US household already has 14.5 connected devices. And the number of global shipments of smart devices is expected to grow to 1.4 billion by 2023. The timing for WiFi 6 couldn’t be better.
Yet WiFi 6 will fall short of its promises without intelligent management and optimization of the home network. For the new era of connectivity to live up to all the excitement, the technology’s limitations need to be addressed. It’s a complex task, but it is within reach, given the right solution.
This article was originally published on EE Times.
Bill McFarland is the CTO of Plume. He leads projects in data science, optimization, standards, intellectual property, and regulatory matters. Previously VP of Technology at Qualcomm, and the CTO of Atheros Communications, Bill holds over 80 patents and has authored over 35 technical papers. He received a Bachelor’s Degree in Electrical Engineering from Stanford University and a Master’s Degree in Electrical Engineering from the University of California. Bill was elected fellow of the IEEE in 2014.