Congestion mitigation in LTE base stations using radio resource allocation techniques with TCP end to end transport

Ghosh, A 2019, Congestion mitigation in LTE base stations using radio resource allocation techniques with TCP end to end transport, Masters by Research, Engineering, RMIT University.


Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Ghosh.pdf Thesis application/pdf 6.07MB
Title Congestion mitigation in LTE base stations using radio resource allocation techniques with TCP end to end transport
Author(s) Ghosh, A
Year 2019
Abstract As of 2019, Long Term Evolution (LTE) is the chosen standard for most mobile and fixed wireless data communication. The next generation of standards known as 5G will encompass the Internet of Things (IoT) which will add more wireless devices to the network. Due to an exponential increase in the number of wireless subscriptions, in the next few years there is also an expected exponential increase in data traffic. Most of these devices will use Transmission Control Protocol (TCP) which is a type of network protocol for delivering internet data to users. Due to its reliability in delivering data payload to users and congestion management, TCP is the most common type of network protocol used. However, the ability for TCP to combat network congestion has certain limitations especially in a wireless network. This is due to wireless networks not being as reliable as fixed line networks for data delivery because of the use of last mile radio interface. LTE uses various error correction techniques for reliable data delivery over the air-interface. These cause other issues such as excessive latency and queuing in the base station leading to degradation in throughput for users and congestion in the network. Traditional methods of dealing with congestion such as tail-drop can be inefficient and cumbersome. Therefore, adequate congestion mitigation mechanisms are required. The LTE standard uses a technique to pre-empt network congestion by a mechanism known as Discard Timer. Additionally, there are other algorithms such as Random Early Detection (RED) that also are used for network congestion mitigation. However, these mechanisms rely on configured parameters and only work well within certain regions of operation. If the parameters are not set correctly then the TCP links can experience congestion collapse.

In this thesis, the limitations of using existing LTE congestion mitigation mechanisms such as Discard Timer and RED have been explored. A different mechanism to analyse the effects of using control theory for congestion mitigation has been developed. Finally, congestion mitigation in LTE networks has been addresses using radio resource allocation techniques with non-cooperative game theory being an underlying mathematical framework. In doing so, two key end-to-end performance measurements considered for measuring congestion for the game theoretic models were identified which were the total end-to-end delay and the overall throughput of each individual TCP link. An end to end wireless simulator model with the radio access network using LTE and a TCP based backbone to the end server was developed using MATLAB. This simulator was used as a baseline for testing each of the congestion mitigation mechanisms. This thesis also provides a comparison and performance evaluation between the congestion mitigation models developed using existing techniques (such as Discard Timer and RED), control theory and game theory. As of 2019, Long Term Evolution (LTE) is the chosen standard for most mobile and fixed wireless data communication. The next generation of standards known as 5G will encompass the Internet of Things (IoT) which will add more wireless devices to the network. Due to an exponential increase in the number of wireless subscriptions, in the next few years there is also an expected exponential increase in data traffic. Most of these devices will use Transmission Control Protocol (TCP) which is a type of network protocol for delivering internet data to users. Due to its reliability in delivering data payload to users and congestion management, TCP is the most common type of network protocol used. However, the ability for TCP to combat network congestion has certain limitations especially in a wireless network. This is due to wireless networks not being as reliable as fixed line networks for data delivery because of the use of last mile radio interface. LTE uses various error correction techniques for reliable data delivery over the air-interface. These cause other issues such as excessive latency and queuing in the base station leading to degradation in throughput for users and congestion in the network. Traditional methods of dealing with congestion such as tail-drop can be inefficient and cumbersome. Therefore, adequate congestion mitigation mechanisms are required. The LTE standard uses a technique to pre-empt network congestion by a mechanism known as Discard Timer. Additionally, there are other algorithms such as Random Early Detection (RED) that also are used for network congestion mitigation. However, these mechanisms rely on configured parameters and only work well within certain regions of operation. If the parameters are not set correctly then the TCP links can experience congestion collapse.

In this thesis, the limitations of using existing LTE congestion mitigation mechanisms such as Discard Timer and RED have been explored. A different mechanism to analyse the effects of using control theory for congestion mitigation has been developed. Finally, congestion mitigation in LTE networks has been addresses using radio resource allocation techniques with non-cooperative game theory being an underlying mathematical framework. In doing so, two key end-to-end performance measurements considered for measuring congestion for the game theoretic models were identified which were the total end-to-end delay and the overall throughput of each individual TCP link. An end to end wireless simulator model with the radio access network using LTE and a TCP based backbone to the end server was developed using MATLAB. This simulator was used as a baseline for testing each of the congestion mitigation mechanisms. This thesis also provides a comparison and performance evaluation between the congestion mitigation models developed using existing techniques (such as Discard Timer and RED), control theory and game theory.
Degree Masters by Research
Institution RMIT University
School, Department or Centre Engineering
Subjects Wireless Communications
Keyword(s) LTE
TCP
Game Theory
Congestion
Base Station
Versions
Version Filter Type
Access Statistics: 120 Abstract Views, 62 File Downloads  -  Detailed Statistics
Created: Mon, 30 Mar 2020, 12:50:20 EST by Keely Chapman
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us