EEMBC, the Embedded Microprocessor Benchmark Consortium, is developing new benchmarks for the Internet of Things (IoT) industry.
The first benchmark, intended to ensure optimum efficiency of security implementations, will aim to provide an industry-created and standardised tool for application developers to quickly, accurately and equitably compare the performance and energy efficiency of security solutions targeted at IoT end-point applications.
“Security is a priority of application developers, though they are typically concerned that implementing security functions within their IoT devices will hurt performance and lower battery life,” said EEMBC president, Markus Levy. “Therefore, a critical goal of our new benchmark will be to quantify the latency and energy impact of implementing security to allow developers to select the optimal combination of microcontroller, hardware and/or software security products for their application.”
The EEMBC IoT Security benchmark will be rolled out in phases. The first phase, planned for beta release in the first quarter of 2017, will contain tests to evaluate a variety of standalone security functions such as SHA256, AES and ECC. The benchmark will enable EEMBC members to implement these algorithms on their microcontrollers using software libraries or hardware accelerators, whichever yields the best balance of performance, energy and cost.
In subsequent phases, the working group plans to combine the standalone functions into specific IoT profiles that will allow users to better see, control and optimise the impact of security at the system level. The group’s modular approach also makes it easier for users to take advantage of optimised security implementations, such as cryptographic processors or modules.
The second benchmark is being developed to measure how gateways perform in specific IoT vertical markets. In general, an IoT gateway lives at or near the ‘edge’. It is used to gather and manage data from multiple sources (sensors, IoT devices), process data locally (instead of in the cloud), react to and predict events, and send data to the cloud. IoT gateways are available in many form factors to support a wide range of vertical applications including industrial automation, transportation (fleet management), digital media (retail advertising), smart cities, home automation, agriculture and healthcare.
The IoT gateway benchmark will utilise a distributed approach with client-server interactions and workloads generated across multiple physical ports. The benefit of this methodology is that it will test the system as a whole, including the processor, physical and wireless interfaces (e.g. Wi-Fi, Bluetooth), operating system and other elements.
“The EEMBC IoT gateway benchmark will standardise assumptions about gateway operational conditions to ensure meaningful comparisons between gateway products,” said Paul Teich, principal analyst at Tirias Research and technical advisor to EEMBC. “Today, without this standardised methodology, IoT gateway benchmarking is not realistic, with buyers having to guess about each gateway’s potential performance for things such as sensor fusion, type of processing workloads, and how much data traffic to manage.”
The aforementioned benchmarks will complement the EEMBC Connect benchmark, also in development. The latter will provide a method to reliably determine the combined energy consumption of the system, taking into consideration the real-world effects of sensor inputs and communication (e.g. Bluetooth and Wi-Fi).
For more information visit www.eembc.org