Bandwidth allocation and computation offloading for service specific IoT edge devices

As we witness the advent of Internet of Things (IoT) devices in the forms of connected appliances, household devices and wearables, the augmented computation demands require scrupulous design to ensure high quality of service. Pushing the intelligence to the edge of the network requires offloading the computational workload to the cloud servers. Consequently, subduing the bandwidth scarcity in a local IoT network along with the buffer limitation of IoT devices has become a major concern. Existing works on bandwidth allocation do not consider constraints of devices, gateways and applications. In this work, we formulate an optimization function to achieve efficient bandwidth allocation to the connected IoT devices that exploits memory limitation and energy of IoT devices, time sensitivity and data loss tolerance of applications. Simulation results corroborate that our proposed solution achieves improved computation offloading performance compared to state-of-the-art works.