Decentralized Edge-to-Cloud Load-balancing: Service Placement for the Internet of Things

Abstract

The Internet of Things (IoT) has revolutionized everyday life and expanded the scope of smart services to a broad range of domains. In ubiquitous environments, fog computing has emerged leveraging the resources in the edge-to-cloud continuum to improve the quality of service, while reducing the traffic on cloud infrastructure and networks. In such a distributed ecosystem with heterogeneous resources of various sizes and inherent dynamics such as varying service demand over time, managing resources and services is a major challenge. This paper studies two optimization objectives and formulates a decentralized load-balancing problem for IoT service placement: (global) IoT workload balance and (local) quality of service, in terms of minimizing the cost of deadline violation, service deployment, and unhosted services. The proposed solution, EPOS Fog, introduces a decentralized multi-agent system for collective learning that utilizes edge-to-cloud nodes to jointly balance the input workload across the network and minimize the costs involved in service execution. The agents locally generate possible assignments of requests to resources and then cooperatively select an assignment such that their combination maximizes edge utilization while minimizes service execution cost. Extensive experimental evaluation with realistic Google cluster workloads on various networks demonstrates the superior performance of EPOS Fog in terms of workload balance and quality of service, compared to approaches such as First Fit and exclusively Cloud-based. The findings demonstrate how distributed computational resources on the edge can be utilized more cost-effectively by harvesting collective intelligence.

Efficient Data as a Service in Fog Computing: An Adaptive Multi-Agent Based Approach

“…However, the only QoS metric considered is the percentage of completed tasks. A decentralized, cooperative agent-based load balancing mechanism for service placement in IoT networks is proposed in [25]. Agents locally generate possible assignments of tasks to available resources, that are then globally optimized to maximize edge utilization and minimize execution costs.…”