A Data Traffic Remapping Scheme for Delay Sensitive Industrial Applications

Network topology configuration in an Edge Infrastructure-as-a-Service (EIaaS) paradigm consists of Edge Computational Nodes (ECNs) placement at optimal locations and balanced assignment of IoT nodes among ECNs. The existing ECN placement and mapping schemes perform the IoT-ECN mapping based on the historic/fixed/single-instance data traffic of IoT nodes. These schemes suffer from issues such as increased delay, unbalanced data traffic distribution etc., when data traffic pattern deviates from the initial settings. This work proposes an extension of our previous work, named as Extended-Cost Aware Edge Computational Node Placement (EcoECNP) scheme, to address the mentioned issues under varying (query-based) data traffic scenarios. The scheme utilizes the common nodes (received from the mapping module of coECNP) and their traffic (obtained from requested data traffic) for remapping among ECNs to achieve balanced data traffic distribution among ECNs and minimize the delay. The performance evaluation indicates that the proposed scheme reduces the average delay per IoT node by 10%-53% and achieves 25%-68% more balanced data traffic distribution among ECNs than state-of-the-art schemes under varying data traffic scenarios.