A software defined networking solution for on-demand base station wake-up in softwarized LTE systems
Tutor / director / evaluatorSeyfettin Demirkol, Ilker
Document typeMaster thesis
Rights accessOpen Access
LTE is a standard offering high speed wireless communication, and it is a major milestone towards flourishing the capabilities of 5th generation (5G) mobile networks to meet the exploding traffic growth in the current and future Internet. However, the introduction of new devices and smartphones is creating many challenges to the LTE operators, and one of the most significant challenges is to control the load on the LTE eNB base stations. In the meantime, operators look for ways to reduce their energy consumptions, for which putting the no- or low-throughput base stations to sleep is a promising solution. Therefore, this project is designed to provide a solution, which would detect congestion in the Radio Access part of LTE network and remotely wake up base stations in the congestion zones. If a congestion is detected at one eNB, a nearby eNB that had been put to sleep (for reasons of energy conservation, reduction of interferences, etc.) is woken up. We provide two implementations of this Congestion-aware On-demand eNB Wake-up (COEW) solution: a deployment of traditional LTE network using a softwarized LTE implementation, consisting of EPC and eNB implementations on different machines. The second implementation is based on Software Defined Networking (SDN), which is then integrated to the traditional network setup. For this, Open vSwitch (OVS) is used on eNBs and EPC machines, in order to have the congestion information to be collected by the SDN controller using OpenFlow protocol. SDN controller (based on OpenDayLight-ODL) communicates with COEW Application via Northbound plugins. COEW application then analyzes the data from each eNB and wakes up sleeping eNB remotely. Validation of the two COEW solutions have been done through physical experiments using commercial UEs and aforementioned community-driven SDN software components.