EEMBC announces the release of IoTMark-Wi-Fi, a benchmark designed specifically to measure the power consumption of low-power 802.11 Wi-Fi chips under realistic load situations. EEMBC is a California-headquartered industry consortium that develops benchmarks for embedded hardware and software.
Short-range protocols, which are optimized for low power consumption, have largely dominated the wireless communication for IoT devices. But with recent improvements in chip design have made the more familiar 802.11 protocol viable for IoT. It drastically reduces power usage while keeps its advantages of greater range and reliability, and broader industry support. However, to bring true Wi-Fi to IoT sensors and other edge-nodes, the industry needs to benchmark their power consumption that yields results that are reliable and relevant in real-world situations.
EEMBC addresses this need by basing IoTMark-Wi-Fi on behavioral profile that measures power usage under a carefully selected workload, reflecting the sometimes unexpected demands experienced by IoT devices. “We had to be very clear about what we’re measuring, and why,” says EEMBC President Peter Torelli. “Long battery life is crucial when you’ve got an office or production floor full of IoT sensors, but actual power consumption depends on a wide range of variables.”
Wi-Fi routers can vary widely, for example, in how frequently they address devices on the network, with “noisier” routers causing them to rapidly use up battery life, especially if there are several routers within range. The 2.4 GHz band where 802.11 operates is crowded with non-IoT devices as well, which can further deplete battery life with their own communications. However, this variability makes a single metric, like standby power consumption, unrealistic for most IoT devices. Instead, EEMBC worked with several member companies, including Dialog Semiconductor, Silicon Labs, Infineon, and STMicroelectronics, to develop a behavioral model that takes these variations into account. The benchmark can illustrate this sensitivity even more dramatically when used in conjunction with an RF isolation chamber, such as those made by octoScope.
IoT devices spend the vast majority of their time in standby, ultimately consuming more power in this state than in active communication—even when using low-power protocols like MQTT. For this reason, IoTMark-Wi-Fi’s behavioral profile was designed to include both a “connected standby” and an application layer component. The official score provided by the benchmark reflects a device’s average power consumption over these two components; its parameters can also be adjusted to give a more accurate result for specific, known situations.
“For customers to make informed decisions, the industry needs a way to make objective measurements of battery life for IoT products,” says Omer Cheema, the Senior Director of IoT Wi-Fi Business Unit at Dialog Semiconductor. “That’s why Dialog Semiconductor has been so excited to collaborate with EEMBC in developing a standardized low power Wi-Fi benchmark.”
Perhaps the most user-friendly aspect of the benchmark is the way its output is scaled: running it provides a number that roughly indicates the number of days the device will last on a pair of standard AA alkaline batteries. “Benchmarks can be very technical, abstract measurements, with good reason,” explains Torelli, “so we thought it’d be nice to tie the output to something everyone can understand.”
