Facial reconstruction

Search LJMU Research Online

Browse Repository | Browse E-Theses

Data Traffic Modelling in Mobile Networks for Heterogeneous Types of IoT Services

Dighriri, M (2020) Data Traffic Modelling in Mobile Networks for Heterogeneous Types of IoT Services. Doctoral thesis, Liverpool John Moores University.

[img] Text
Mohammed Dighriri -Thesis.pdf - Published Version
Restricted to Repository staff only until 13 January 2022.

Download (6MB)

Abstract

The upcoming 5th Generation (5G) mobile networks will be different from the previous mobile network generations in the fact that it will enable the mobile networks industry, besides offering superior broadband services, to enhance Internet of Things (IoT) industries such as vehicular communication system, factory automation, smart healthcare system and many more. Many of these use cases have challenging and quite often contradicting requirements in terms of data rate, latency, throughput and so on. This suggests that 5G mobile networks need to adopt flexible models that can adapt to different IoT device and traffic requirements. Consequently, a fresh look into how mobile networks are currently designed and deployed is needed. Historically, mobile networks have relied on the axiomatic role of cells as the cornerstone of the Radio Access Networks (RAN). Mobile network systems have witnessed several recent trends such as the increased heterogeneity in heterogeneous types of IoT services infrastructure and spectrum as well as the rise of different traffic types with different Quality of Services (QoS) requirements. In this direction, this thesis focuses on improving the performance of cell-edge users or IoT devices in 5G mobile networks by initially implementing the network slicing management approach, particularly as, with the fast growth of IoT, billions of devices will join the internet in the next few years. Hence, the latest 5G mobile technologies expected to offer massive connectivity and management ability of high volume of data traffic at the presence of immense interferences from a mobile network of IoT devices. Further, it will face challenges due to congestion and overload of data traffic due to a humongous number of IoT devices. Besides, these devices likely to demand high throughput, low latency and high level of reliability especially for critical real-time smart systems in density and small zone, such as in Vehicular Communication System (VCS), these vehicles mainly rely on connectivity aspects. Furthermore, IoT devices transmit small and large-sized packets with different radio resource requirements. For example, Smart Healthcare System (SHS) devices transmit small-sized of a data with utilizing a small portion of Physical Resource Block (PRB) as the smallest radio resource unit, which is allocated to a single device for data transmission in 5G mobile networks. In the IoT services with transmitting a small-sized data, the capacity of the PRB is not fully utilized, which causes wastage and unfairness of using PRB among these IoT devices or services. The novelties made in this thesis significantly advance a Slice Allocation Management (SAM) model based on critical services such as (VCS) to satisfy low latency demand. The proposed model aims at providing dedicated slices based on service requirements such as expected low latency for (VCS). To ensure such performance to data traffic of IoT devices in Uplink (UL)of Relay Node (RN) cells in the 5G mobile networks by slicing the RAN, along with assigning the nearest Mobile Edge Computing (MEC) with isolating slices depend on technical and QoS requirements for each IoT nodes. Also, this thesis proposes a Data Traffic Aggregation (DTA) model for efficient utilization of the smallest untie of PRB by aggregating the data traffics of several IoT devices, which can support IoT node throughput such as SHS. Also, this thesis presents a comprehensive comparison of the packet scheduling mechanisms include Priority Queuing (PQ), First-In-First-Out (FIFO) and Weighted Fair Queuing (WFQ) applied based on data traffic slicing model through RN cells. These thesis models are validated through the OPNET simulator to measure the performance of the SAM and DTA Models along with the assessment of packet scheduling mechanism. The simulation considers IoT devices in various smart systems such as VCS, SHS and smartphones also, different protocols include Simple Mail Transfer Protocol (SMTP), File Transfer Protocol (FTP), and Voice over Internet Protocol (VoIP) and Real-time Transport Protocol (RTP). Simulation results show a significant improvement in IoT nodes packets transmission via RNs and Donor eNodeB (DeNB) cells, in My SAM Model scenario comparing with other scenarios. The model has improved such as End-to-End (E2E) delay in FTP node by reaching 1ms, loading in VoIP node by 80% and throughput of all nodes in the uplink side of networks by 66%. In addition, the results display significant impact of IoT data traffic with different priority, networks E2E performance is improved by aggregating data traffic of several IoT devices with DTA model, which is determined by simulating several scenarios, considerable performance improvement is achieved in terms of averages cell throughput, upload response time, packet E2E delay and radio resource utilization. Finally, the result found PQ packet scheduling mechanism as the appropriate scheduling mechanism in case of supporting several of priorities queuing for data traffic.

Item Type: Thesis (Doctoral)
Uncontrolled Keywords: 5G mobile networks; IoT
Subjects: Q Science > QA Mathematics > QA76 Computer software
Divisions: Computer Science
Date Deposited: 20 Jan 2020 10:34
Last Modified: 20 Jan 2020 10:34
DOI or Identification number: 10.24377/LJMU.t.00011950
Supervisors: Lee, GM, Shamsa, TB and Pereira, R
URI: http://researchonline.ljmu.ac.uk/id/eprint/11950

Actions (login required)

View Item View Item