This paper presents an implementation of a distributed TinyML system and an analysis of the network overhead in communication between devices. We use a pipelined approach, where each device computes a portion of the layers and sends intermediates to the next device. Our results show that the communication overhead strongly dominates, taking over five times longer than computation time. We present possible solutions to mitigate this overhead, including smart partitioning of models, optimizing device placement and signal strength, and alternate wireless protocols such as WiFi. These strategies aim to make the deployment of several TinyML devices working together a more practical scenario.

Edge AI computing boxes are a new class of computing devices that are aimed to revolutionize the AI industry. These compact and robust hardware units bring the power of AI processing directly to the source of data-on the edge of the network. On the other hand, on-demand serverless inference services are becoming more and more popular as they minimize the infrastructural cost associated with hosting and running DNN models for small to medium-sized businesses. However, these computing devices are still constrained in terms of resource availability. As such, the service providers need to load and unload models efficiently in order to meet the growing demand. In this paper, we introduce FusedInf to efficiently swap DNN models for on-demand serverless inference services on the edge. FusedInf combines multiple models into a single Direct Acyclic Graph (DAG) to efficiently load the models into the GPU memory and make execution faster. Our evaluation of popular DNN models showed that creating a single DAG can make the execution of the models up to 14% faster while reducing the memory requirement by up to 17%. The prototype implementation is available at https://github.com/SifatTaj/FusedInf.

To cope with the high resource (network and compute) demands of real-time video analytics pipelines, recent systems have relied on frame filtering. However, filtering has typically been done with neural networks running on edge/backend servers that are expensive to operate. This paper investigates on-camera filtering, which moves filtering to the beginning of the pipeline. Unfortunately, we find that commodity cameras have limited compute resources that only permit filtering via frame differencing based on low-level video features. Used incorrectly, such techniques can lead to unacceptable drops in query accuracy. To overcome this, we built Reducto, a system that dynamically adapts filtering decisions according to the time-varying correlation between feature type, filtering threshold, query accuracy, and video content. Experiments with a variety of videos and queries show that Reducto achieves significant (51-97% of frames) filtering benefits, while consistently meeting the desired accuracy.