Traffic Modells for Voice over LTE

OData support
Supervisor:
Dr. Molnár Sándor
Department of Telecommunications and Media Informatics

Telecommunication providers are striving for introducing VoLTE service all over the world. This service provides the transfer of calls over 4G network, besides it aims among others to warrant higher quality and cause lower burden for the mobile equipment and thereby for the subscriber than the traditional voice transmission on mobile device. It originates the theoretical opportunity for seceding from the previous 2G and 3G networks and finishes the realization of circuit switched services on packet switched network.

I became acquainted with the indispensable entities of LTE and IMS network through the VoLTE service and the monitoring system thanks to AITIA International Inc, which is continuously tracking the LTE network of Magyar Telekom Plc.

After I had got to know the LTE specific identifiers and the Attach procedure, I was able to assemble the S1AP CDRs going through S1-MME interface of LTE network and decipher the security and integrity protected messages relating to the mentioned interface.

I investigated the real-time deciphering system of Magyar Telekom Plc.’s LTE network traffic, developed by AITIA International Inc by means of the mentioned monitoring system and shed light on some problematic mechanisms of the deciphering system.

I received real SIP traffic by AITIA International Inc. It was processed by my Python program and therefore a time series from SIP CDR’s opening times was established, which provided the base of network traffic analytics and modelling.

I examined the stationarity of this time series by my MATLAB function and omitted some of its elements in order to warrant the stationarity of the remained time series, on which the analysis went on. I determined and illustrated the main descriptors of the time series by built in and my own MATLAB functions, respectively.

I created a Poisson model appropriately for the received results and verified by a single queueing performance simulation test in MATLAB. I compared the behaviour of this queue in case of the measured traffic and the generated traffic fitting on my model. According to the validation test, the constructed Poisson model describes the measured network traffic well.

Downloads

Please sign in to download the files of this thesis.