IoT Query Latency Enhancement by Resource-Aware Task Placement in the Fog

The advancement of the IoT (Internet of Things) domain has led to the widespread adoption of IoT applications. These latency-sensitive IoT applications demand low query latency for real-time data analytics. Fog computing has aided in mitigating query response time regarding single query optimization. However, challenges exist regarding optimizing concurrent execution of simultaneously arriving queries within the heterogeneous and resource-constrained fog environment. This paper presents an efficient resource-aware multi-query optimization technique to address these challenges. The proposed technique formulates a query execution plan in a network-compute-aware manner to optimize the concurrent execution of multiple queries. We introduce novel network-compute-aware multi-query task placement algorithms for narrow and wide transformation tasks that take into account the current network and computational resource statistics. The proposed algorithms perform optimal task placement to reduce the overall latency of multiple queries in a resource-aware manner. Our evaluation reveals that our technique can reduce the query latency by up to 76%, and decrease the network usage by 60% when compared to other techniques.