DENIS-SDN: Software-Defined Network Slicing Solution for Dense and Ultra-Dense IoT Networks

TL;DR

This paper introduces DENIS-SDN, an open-source SDN-based solution for dense IoT networks that improves packet delivery ratio through network slicing, addressing interference, congestion, and QoS challenges.

Contribution

It presents a novel SDN framework with network slicing and density control mechanisms tailored for ultra-dense IoT environments, enhancing network performance.

Findings

Up to 4.8% improvement in PDR with logical slicing.

Up to 11.6% improvement in PDR with physical slicing.

Effective management of interference and congestion in dense IoT networks.

Abstract

Traditional Wireless Sensor Networks protocols used in Internet of Things Networks (IoTNs) today face challenges in high- and ultra-density network topology conditions. New networking paradigms like Software-Defined Networks (SDN) have emerged as an up-and-coming approach to address IoT application requirements through implementing global protocol strategies and network programmability. This paper proposes a divide-and-conquer solution that aims to improve the PDR in ultra-dense IoT (UDIoT) network environments using network slicing. As such, we develop and evaluate DENIS-SDN, an open-source SDN solution for UDIoT Network environments consisting of a modular SDN controller and an OpenFlow-like data-plane protocol. DENIS-SDN utilizes our Network Density Control mechanism based on operational specification requirements, which address the challenges UDIoT network deployments pose, including interference, congestion, resource management, control, and quality of service (QoS) performance issues. To achieve this, it divides dense IoT networks into either logically sliced sub-networks separating nodes using routing rules or physically sliced sub-networks separating nodes into different radio channels. We provide evaluation results over realistic scenarios demonstrating improved PDR performance up to 4.8% for logically and up to 11.6% for physically sliced network scenarios.