SDN-based scheduling strategy on load balancing of virtual sensor resources in sensor-cloud

摘要

Sensor-cloud has recently emerged as a promising technology that provides sensor management approaches and facilitates sensor data sharing by virtualizing the physical sensors on a cloud platform. This framework is a new trend to deploy and execute various smart applications with rapidly changing in the amount of traffic loads. Therefore, it is essential to provide a platform to meet demand variation, while ensuring a given QoS. In this paper, we present a scheduling strategy on load balancing of virtual sensor resources based on Software Defined Networking (SDN). To do that, we propose to model system using queuing model with variable number of virtual sensor instances. Experimental results prove that this method is able to realize load balancing and reasonable resources utilization both in stable and variant system load.