A P4-assisted task offloading scheme for Fog networks: An intelligent transportation system scenario

In Fog-based Internet of Things (IoT) networks, Fog computing offers great promises to achieve efficient task processing among IoT devices and the Cloud with lower processing delay and a lighter bandwidth burden. However, the sustainability of these networks is critical so that Fog servers efficiently allocate/offload increasing task types to appropriate resources. In this paper, a novel task offloading scheme called Task Offloading Scheme with P4 (TOS-P4) is proposed for Fog-based IoT networks using Programming Protocol-independent Packet Processors (P4) technology. The proposed scheme is evaluated through an Intelligent Transportation System (ITS) application scenario and compared to a conventional model called Task Offloading Scheme with Software-Defined Networking controller (TOS-SDN). In this scenario, the network traffic is regulated by P4, allowing task offload decisions to be more flexible and programmable according to server loads and predetermined forwarding rules. We present the applicability of the proposed task offloading scheme in the Fog-based IoT network and its benefits to network performance in terms of latency and computation overhead. According to the experimental results, when the servers' load status is measured at 5 s intervals, TOS-P4 is 6.54 times more efficient than TOS-SDN in waiting times of tasks received at Resource Poor (RP) Fog servers. Also, in TOS-SDN case, the average waiting time of a task on RP Fog servers is 30 times more compared to Resource Rich (RR) Fog servers.