Closed mtrifiro closed 5 years ago
Suggest we add: "Although tasks requiring no-latency or minimal latency processing are performed at the edge device (e.g. sensor) where possible, these devices are highly constrained in terms of energy consumption and cost."
An edge computing use case where tasks are offloaded from an edge device to the infrastructure edge for remote processing. Computational offloading seeks, for example, performance improvements and energy savings for mobile devices by offloading computation to the infrastructure edge with the goal of minimizing balancing task execution latency and with mobile device energy consumption and cost. Computational By offloading work from the device, developers can build new classes of mobile applications that would require computational power and storage capacity that exceeds what the device alone is capable of employing (e.g., untethered Virtual Reality). Many classes of devices, from robotic drones to self-driving cars and remote sensors, will benefit from the lower costs and reduced battery consumption possible with offloading. In other cases, workloads may be offloaded from a centralized to an edge data center for performance. The term is also referred to as cloud offload and cyber foraging in the literature.
That works, will add - though I still wish it weren't called cyber foraging. 👍
Make the point that edge computing near the sensor is preferred for latency concerns for a wide range of applications including robotics/drones, self-driving cars, IoT, and portable/wearable electronics. However, at the sensor there are often stringent constraints on energy consumption and cost.