Software-Defined Networking (SDN) paradigm provides the ability to handle mobility more efficiently due to its programmability and fine granularity. However, in this emerging setting, the handover procedure still suffers delay due to exchanging and processing handover signaling messages. In this paper, we study the relevancy between an SDN controller’s load and handover delay. We show that an over-loading state can prolong handover delay, so as a countermeasure, reaching that state is mitigated by applying a load balancing mechanism. Our primary metric is the controller’s response time, as it directly affects the completion of any mobility-related procedure. We propose a load balancing management framework that deploys two concepts: network heterogeneity and context-aware vertical mobility. Our proposal is composed of three main aspects. First, we identify candidate users based on their context information. Second, we reduce the frequency of load dissemination between multiple controllers, and hence, reducing processing and communication overhead. Third, after the candidate users are determined, we optimize the decision problem on the selection among heterogeneous candidate networks. Through simulation, our framework has shown as much drop as a 28% drop in response time compared to previous proposals.

由于软件定义网络（SDN）的可编程性和精细粒度，它提供了更有效地处理移动性的能力。然而，在这种新兴环境中，由于交换和处理切换信令消息，切换过程仍然遭受延迟。在本文中，我们研究了SDN控制器的负载与切换延迟之间的相关性。我们显示，过载状态可以延长切换延迟，因此，作为对策，可以通过应用负载平衡机制来缓解达到该状态的情况。我们的主要指标是控制器的响应时间，因为它直接影响任何与移动性相关的过程的完成。我们提出了一个负载平衡管理框架，该框架部署了两个概念：网络异构性和上下文感知的垂直移动性。我们的建议包括三个主要方面。首先，我们根据候选用户的上下文信息来识别他们。其次，我们减少了多个控制器之间的负载分配频率，因此减少了处理和通信开销。第三，确定候选用户后，我们优化了异构候选网络中选择的决策问题。通过仿真，与以前的建议相比，我们的框架显示响应时间最多下降了28％。