A Dynamic Multi-Resource Management for Edge Computing

Abstract

With IoT devices are widely deployed, numerous IoT data are generated and transmitted to cloud servers for further process. As numerous IoT data usually lead to congestion, the edge computing technology is further emerging to alleviate this problem. Generally, the decision that which data are transmitted to edge nodes or cloud servers and which data stay will significantly influence the system performance. Therefore, proposing a practical resource management for edge computing becomes a necessity. However, in IoT environments, not only the data request model is highly dynamic but also multiple types of resource are required in resource allocation. Existing resource managements cannot totally solve these problems. To this end, this paper proposes a Dynamic Multiple Resource Management (DMRM) applying the Multi-Resource Binary Particle Swarm Optimization (MR-BPSO) to allocate multiple resources in dynamic IoT environments. Moreover, three experiments are provided to present the task complete rate in three situations, including the dynamic request, dynamic resource as well as dynamic request and resource. Comparing with other resource managements, the proposed MR-BPSO has better performance, and therefore the DMRM is more applicable in dynamic IoT environments.