A Study on Round-Trip Time Performance of Tactile Internet using Multilevel Cloud Structure

Abstract

The tactile Internet (TI) is poised to be evolutionary to the Internet of Things (IoT), enabling real-time immersive interactions by transmitting a sense of touch and actuation over remote distance with audio or visual haptic feedback. A few of the stringent requirements of TI are achieving a round-trip time (RTT) of 1 ms, ultra-high reliability and availability. Addressing these challenges, the emerging TI can bring a plethora of applications, such as Industry 5.0 (industry automation), teleoperation frameworks (self-driven vehicles), augmented and virtual reality (AR/VR) related applications and Healthcare 5.0 (telesurgery, telediagnosis and telerehabilitation). Whilst these applications have been promising, TI is still in the nascent stage of development. The potential enablers of TI, such as the fifth-generation (5G) framework, software-defined networking (SDN), edge/fog computing (EC/FC), network slicing (NS), and an appropriate multiple access (MA) scheme, can meet these TI applications’ requirements and their challenges. The ultra-low latency and high data rates of 5G provide a foundation for real-time communication. SDN enables dynamic traffic management and efficient resource allocation, reducing network congestion. EC/FC brings computation closer to end-users, minimising transmission delays. NS allows for isolating TI traffic, ensuring dedicated resources for critical applications. Finally, an MA, specifically the Non-orthogonal Multiple Access (NOMA) scheme, enhances spectral efficiency, supporting multiple users with minimal latency overhead. Collectively, these technologies address core requirements, mitigate round-trip time and enhance the responsiveness of TI systems.

This PhD research aims to study the RTT performance in TI communication infrastructure, employing a multilevel cloud structure with FC and SDN. Moreover, NS and an appropriate MA scheme are incorporated to meet the stringent demands or requirements of TI. Firstly, a comprehensive survey on TI is presented, focussing on design architecture, crucial application areas, current issues and challenges with potential enabling technologies. To emphasise the novelty of the survey, the brainstorming mind map is presented, covering all the discussed topics. In addition, an extensive literature review of the mentioned potential enablers of TI is explored.

Secondly, a novel fog-based traffic flow framework is proposed, employing FC and SDN to address the TI challenges of having an RTT of 1 ms and ultra-high reliability for establishing haptic communications. An effective traffic flow algorithm is developed to route the traffic in a multilevel cloud structure efficiently. The combined SDN and FC approaches with the developed algorithm provide an effective solution to reduce extra processing and waiting times at each level of the cloud-based structure. Hence, the RTT is also reduced when traffic flows from master to slave sections and vice versa. The performance metrics such as throughput, RTT, energy consumption, and reliability are evaluated using the iFogSim simulator. The simulation results for the proposed system outperform those of the existing edge, cloud, and cellular networks.

Thirdly, a novel NS mechanism based on TI communication infrastructure is proposed, leveraging SDN with Open vSwitch (OVS) to resolve the issue of provisioning and controlling the network slices on demand. Hence, the network slicing algorithms are developed for three different scenarios: topology slicing, service slicing, and emergency slicing, to have tailored and customised slices according to network requirements. This communication infrastructure is sliced under diverse traffic such as UDP, TCP and other traffic packets using pre-designed slice configurations. The system performance is evaluated with metrics such as throughput and RTT by Vagrant environment and Mininet. The simulation results align with pre-designed slices’ configurations, i.e., the preset allocated spectrum and RTT constraint values, and show the efficacy of the proposed NS algorithms, thus validating these algorithms with existing algorithms.

Fourthly, a novel downlink power domain Single-Input Single-Output (SISO) NOMA communication scenario for TI is proposed, employing multiple sensors and actuators collectively treated as users and a base station. An analytical system model is mathematically derived under the NOMA scheme, incorporating signal-to-interference and noise ratio (SINR), sum rate, and fair power allocation (PA) coefficients. Considering two-user and three-user scenarios, the system performance is analysed and evaluated bit error rate (BER) and sum rate between NOMA and orthogonal multiple access (OMA) by varying path loss exponent and fixed PA coefficients using MATLAB. Moreover, the outage probability and achievable sum rate are analysed by varying fixed and fair PA coefficients in the NOMA scheme. For higher SNR, the achievable sum rate for the proposed NOMA system outperforms OMA, thus maximising spectral efficiency, minimising latency and promoting dynamic PA and user fairness. Finally, 4 × 4 Multiple-Input Multiple-Output (MIMO) NOMA performance is analysed by incorporating zero forcing-based beamforming and a round-robin scheduling process with SISO-NOMA concerning achievable sum rate and latency.

Lastly, the proposed models in this research have been validated against state-of-the-art models, proving performance superiority, resource efficiency and scalability.