Multi-attribute-based QoS-aware virtual network function placement and service chaining algorithms in smart cities

Mobile Edge Cloud (MEC) and Network Function Virtualization (NFV) have been emerging as promising technologies on providing low-latency IoT services in smart cities. However, the MEC can only provide limited number of IoT services because of its finite computing resource. How to deploy VNF instances and route data packets for IoT services in the MEC has been considered as a nontrivial problem. The problem becomes more complicated when taking Quality of Service (QoS) into consideration. In this paper, we study the VNF placement and service chaining problems while considering multi-attribute-based QoS in smart cities. We first formulate the problem as an Integer Linear Programming (ILP) and utilize randomized rounding technique to calculate a solution to the problem. Considering IoT services have different QoS requirements, we design an unsplittable flow based heuristic algorithm to deploy computing resource for each type of IoT service. Finally, large quantities of simulations prove that our algorithms outperform their counterparts. [Display omitted] •The smart city scenario composed of NFV-enabled mobile edge clouds.•The throughput maximization problem in the mobile edge cloud.•The dynamic multi-attribute QoS assessment mechanism in the mobile edge cloud.•The randomized-rounding-based VNF placement and service chaining algorithm.•The VNF placement algorithm for IoT services with special QoS requirements.