Title: Testing QoE in Live Wireless Networks at the Dawn of 5G Technology
During the last decade, the 4G networks recorded significant multi-folded and intertwined technical changes in the delivery of voice/video services towards an augmented QoE. First, the wireless evolution to all IP technology which enabled IP voice (e.g. VoLTE) delivery as well as the introduction of advanced transmission techniques that enabled adaptive bandwidths allocation for video streaming delivery. Second, the multimedia signal processing advancements which facilitated HD voice and UHD/4K video delivery on wireless access. Third, the variety of smart devices which embed sophisticated clients for an enhanced voice/video QoE. And then, the variety of voice/video services types (e.g. carrier, OTTs) delivered on various (encrypted) protocols and consumed in different scenarios; each diversely contributing to services’ QoE. Today, at the dawn of 5G technology the wireless networks’ and devices’ complexity increase exponentially and multi-dimensionally; new services types (e.g. as AR/VR) become a reality.
The talk briefly discusses how traditional standardized voice/video QoE models coped with 4G related aspects and presents solutions that fill the gaps left open by these models regarding practical engineering implementation in live network testing tools, something crucial for the operators to benefit out of these models. Then, the talk addresses the disruptive aspects that 5G technology brings on QoE evaluation techniques and discusses QoE modeling needs to evolve towards techniques that better capture the new intertwined complexities and new services’ types, as well as how the testing tools have to adapt to cope with operators needs for cost efficient testing solutions and with 5G context aware QoE delivery aspects. The talk describes how applied artificial intelligences and machine learning can be used towards these goals and what are the limitations.
Speaker: Dr. Irina Cotanis
Dr. Irina Cotanis holds an Doctorate in Electrical Engineering and a score card of more than a 30 years of experience in wireless communications systems, statistical signal processing and statistics, as well as 20 years as active member in standardization organizations, and of more than 100 publications, (e.g. IEEE conference presentations, industry seminars, ITU/ETSI standards and student books). She has been awarded international patents and acted as reviewer to IEEE journals and conferences, and as session chair to various IEEE conferences.
Dr. Cotanis started her career as a teaching professor in academia and in late ‘90s she joined the wireless industry working with Ericsson as research expert and then with network testing tool vendors, lately InfoVista, as head of technology. In this position, she put her radiocommunication and statistical signal processing knowledge and analytical skills to model complex technical issues governing wireless networks and to transform them into practical network testing engineering solutions. The technology projects she leads cover: voice/video QoE models implementation, QoE centric network and services quality assurance and predictive analytics techniques for autonomous cloud-based network testing systems. The latest project is focused on the disruptive aspects that 5G technology brings on QoE evaluation techniques within the framework of the 5G context aware QoE services delivery. Within this project, Dr. Cotanis leads research efforts on machine learning based QoE solutions suitable for 5G device, network and service context aware testing and monitoring. Along the same lines, she is also currently investigating predictive analytics techniques to support AI augmented network testing methodologies.