The smart world is connecting all universe more than ever thought possible, benefiting from the significant advances of the Internet of Things (IoT) applications using wireless sensor networks (WSN) as the core technology. A challenging issue in the IoT paradigm is the heterogeneity in different parts of the network. The network developers need to use resources belonging to different platforms for their applications, and the software defined network (SDN) approach is a mainly considered solution. In this paper, a software defined wireless sensor network (SDWSN) with an energy predictor model (SDWSN-EPM) based on the Markov probabilistic model is proposed to reduce the energy consumption and the network latency. The energy consumption rate (ECR) of the sensor nodes is modeled using the Markov model and the states of the sensor nodes. The ECR is used by the SDN controller to predict the residual energy level of the nodes and consequently, the energy consumption of the network. The cumulative distribution functions (CDF) of the delay, power consumption, and the network lifetime in both SDWSN and SDWSN-EPM schemes are compared. The results confirm that the SDWSN-EPM model significantly improves the performance of the sensor networks.

得益于以无线传感器网络（WSN）为核心技术的物联网（IoT）应用的巨大进步，智能世界正在以前所未有的方式连接所有宇宙。物联网范式中的一个挑战性问题是网络不同部分的异构性。网络开发人员需要为其应用程序使用属于不同平台的资源，而软件定义网络（SDN）方法是主要考虑的解决方案。本文提出了一种基于马尔可夫概率模型的带有能量预测模型（SDWSN-EPM）的软件定义的无线传感器网络（SDWSN），以减少能耗和网络延迟。使用马尔可夫模型和传感器节点的状态对传感器节点的能量消耗率（ECR）进行建模。SDN控制器使用ECR预测节点的剩余能量水平，从而预测网络的能量消耗。比较了SDWSN和SDWSN-EPM方案中的延迟，功耗和网络寿命的累积分布函数（CDF）。结果证实，SDWSN-EPM模型显着提高了传感器网络的性能。