Device Management at the Network Edge to Support QoS-aware IoT Services

The Internet of Things (IoT) brings Internet connectivity to everyday devices. The M2M systems generated by these IoT devices produce a large amount of data which is then forwarded to the IoT applications for processing, decision making, etc. Best-effort delivery is sufficient in some circumstances; yet, in others, like in the context of Industrial IoT (IIoT), IoT applications may have strict Quality of Service (QoS) requirements, particularly in terms of latency. The severe energy, memory, processing, and communication constraints of IoT devices and networks, however, make it difficult to enable the deployment of QoS-aware services. To support QoS-aware services provided by IoT devices, we propose providing a set of improved functionalities between IoT devices and IoT applications, which are frequently set up in the cloud. We propose integrating these functions into an edge-based IoT proxy that can efficiently utilize network, computing, and storage resources at the edge and take advantage of the close proximity to IoT devices to enhance their communication with IoT applications. The proposed functions aim (i) to reduce and control the flow of requests sent to the IoT devices, resulting in benefits such as reduced and more predictable end-to-end latency and reduced packet loss; (ii) to minimize the energy consumed by the IoT device for sensing the environment and for transmitting the status update messages, while taking into account the information freshness requirements of the IoT applications; and (iii) to lower the size of the packets to comply with the required maximum packet length of energy-constrained IoT networks. We evaluate the proposed functions both by simulation and experimentally, showing that they can strongly improve the QoS of the IoT system.