Improved IoT Application Placement in Fog Computing Through Postponement

In fog computing, to ensure Quality of Service (QoS), we need to specify a placement plan for distributing IoT applications among computing devices for processing; this is the IoT application placement problem. With a potentially huge number of computing devices and applications, solving the IoT application placement problem can be decentralized, i.e., independent optimization can be performed in parallel across diverse clusters, thus mitigating the networking and computing overhead and enhancing the QoS consequently. However, deploying applications in clusters based on available resources at the time of placement decision may lead to the propagation of undeployed applications among clusters, in which case ensuring the QoS constraints of those applications is not guaranteed. This paper proposes an algorithm that can be integrated into any existing placement algorithm to enhance the QoS of undeployed applications through postponement rather than immediate propagation among clusters. Initial experimental results from integrating this algorithm into an existing placement algorithm show an improvement in the percentage of applications satisfying QoS.