Understanding the characteristics of network traffic patterns in cellular radio access systems is vital to aspects ranging from conception, design, implementation, testing, and rolling out new network features and concepts. Over the years the main observation was that the overall 4G-user behavior tends not to change radically over long periods of time. With 5G it is foreseen that the traffic behavior will drastically change, mainly due to new service-tailored features.
Live networks are inherently dynamic environments where several traffic generating applications evolve and run side-by-side, subjecting the user behavior and the application mix to a constant shift as the technology evolves. With 5G the traffic mix complexity will be further altered, and a legacy 4G-like application aggregated view is deemed to no longer provide the adequate means for development of robust and sustainable solutions.
The practical aspects of the charaterisation of an observed traffic behavior manifest itself in product develpoment in what is known as traffic models on various levels, complexity and resolution. The development of advanced traffic models using pertinent mix of application-level traffic models is thus necessary in order to offer a better resolution for realistic tests, design relevant input, and prediction of future 5G network usage. The proposed thesis adopts the use of machine learning techniques to unearth the application-level network traffic behavior and group it into independent traffic models, which in their turn, when combined in a controlled manner, will generate a representative, fine-tuned and scenario dependent aggregated user behavior at network level.
Generating application-level traffic models is a non-trivial task as there is an unknown mix of device types, user behaviors, application behaviors, device configurations, etc. In accordance to traditional traffic modelling studies, real-life network data from customer measurement campaigns shall be used, but advanced analytic methodologies shall be employed to utilize the data’s full potentials.
Based on data containing basic labeling of targeted applications/services, machine learning shall be exploited to:
- Identify the individual behaviors of conceivably all applications/services. (ambition level 1)
- Several applications/services tend to present a variety of complex behaviors. For those applications/services, the most characteristic behaviors should be identified and grouped using statistics.
- Categorize the identified application/services based on similarities in their behavior in a limited number of groups. (ambition level 1)
- Infer the possible correlations between co-deployed applications/services, different user behaviors, and device or service types. (ambition level 2)
Part of the thesis work will be to setup a proper tool chain for the anlysis work, that can be exploited by Ericsson for further analysis needs, outside the scope of the current thesis. Components used today for analysis work includes scripting in Linux/Unix, commercial Statistical Analysis System (SAS), MATLAB, and Python. These components can be seen as the basis for development work, but not necessarily a limitation.
The thesis will be concluded with a result presentation for the Ericsson research team.
Qualifications: This project is targeting students in electrical engineering, computer science, computer engineering or similar fields. Background in wireless communication is preferred.
Extent: 1 student, 30hp
Location: Ericsson AB Mjärdevi, Linköping
Preferred Starting Date: Spring 2020
Keywords: Machine learning, Mobile Telecommunication, Network traffic Modeling, 5G applications
Last day of application 31st of October 2019
For any queries please contact recruiter Sarah Lashari, email@example.com
Ericsson provides equal employment opportunities (EEO) to all employees and applicants for employment without regard to race, color, religion, sex, sexual orientation, marital status, pregnancy, parental status, national origin, ethnic background, age, disability, political opinion, social status, veteran status, union membership or genetics.
Ericsson complies with applicable country, state and all local laws governing nondiscrimination in employment in every location across the world in which the company has facilities. In addition, Ericsson supports the UN Guiding Principles for Business and Human Rights and the United Nations Global Compact.
This policy applies to all terms and conditions of employment, including recruiting, hiring, placement, promotion, termination, layoff, recall, transfer, leaves of absence, compensation, training and development.
Ericsson expressly prohibits any form of workplace harassment based on race, color, religion, sex, sexual orientation, marital status, pregnancy, parental status, national origin, ethnic background, age, disability, political opinion, social status, veteran status, union membership or genetic information.
Primary country and city: Sweden (SE) || || Linköping || R&D
Req ID: 300269