Forensic-Aware Load Balancing for Fog-Enabled IoT Resources Using Hierarchical Clustering Analysis

Abstract

Due to increasing amounts of apps using IoT, there has been a focus on properly managing resources, having low latency communication, and processing real-time data across distributed computing environments. While cloud computing allows for scalability of storage and processing power, it also introduces potential delays as a result of its architecture in terms of requirements for bandwidth for latency-dependent applications. Fog computing expands cloud-based functions to the edges of the network, thus providing distributed computing and networking capabilities. In the case of heterogeneous devices, the main problems relate to workload management, resource management, and forensic preparedness. This paper presents a resource load-balancing architecture that supports forensic awareness for fog OS-based IoT resources. Hierarchical clustering is implemented to provide an optimized load balancing system to monitor the resources and distribute workload. This data workload classification scheme identifies the various categories of data based on their processing requirement and allocates resources accordingly. The intelligent data migration framework will be used with all data processed in order to have a forensic awareness about any data that would be affecting the performance of the system, through the systematic tracking of workloads and resource management, allowing evaluation of this framework and its implementation to be done in the iFogSim simulation environment using various configurations of the network. Analysis of test samples shows that, on average, newer designs will have marked improvements over traditional structures (cloud-based) regarding latency, execution costs, network usage, energy use, and elapsed time. This same analysis also supports using hierarchical clusters and using a fog-cloud-based forensic-aware load balancer type with these hierarchical clusters to have an elasticity mechanism for resource optimization and to minimize response times to requests, which will maximize a system’s ability to provide visibility and to use IoT-enabled devices in real-time applications.