Read the interview with Felicián Németh and his winning proposal: QUIC to Enhance SPIRIT’s Telepresence (QUEST)
Felicián Németh and HUN-REN TKI
The HUN-REN Office for Supported Research Groups (TKI) is part of the Hungarian Research Network, which comprises of many research centres, research institutes, and research groups operating at universities, conducting basic and applied research, exploring the most varied disciplines of mathematics and physical sciences, natural sciences, life sciences, social sciences and the humanities. The research group, which F. Németh and his team are the members of, has been focusing on network softwarisation and cloud technologies. The team has been conducting active research in the fields of Software Defined Networking, cloud native technologies, 5G/6G systems and related business aspects, future immersive services including cloud-based XR applications.
Can you give a brief overview of your winning proposal?
What are its key objectives and innovative aspects?
QUIC is a recently standardised general transmission protocol that has many favourable properties for real-time media transmission over the de-facto WebRTC protocol suite, which is used throughout the SPIRIT project. In resource-constrained situations, for example, when there is a sharp bandwidth drop or right after a client migration event, QUIC can provide better quality by omitting the transmission of less relevant data, compared to randomly dropping data packets. Several proposals exist for achieving prioritisation by distributing the video in QUIC streams. However, finding the best-performing scheduling strategy for telepresence applications requires further research.
We’ll extend our existing QUIC-based video transmission toolset with multi-streamed video transmission capability; integrate our GStreamer-based sender with the holographic telepresence scenario; study and implement different stream scheduling strategies to improve QoE; extend our existing measurement tool to receive multi-streamed video transmission; conduct measurements in imec’s testbed; and disseminate the results.
What motivated you to apply for the SPIRIT Open Call?
Our research group is always on the lookout for interesting international collaboration options that are related to our research fields but at the same time carry the possibility for exploring new technologies, concepts or applications. SPIRIT proved to be just the right choice for achieving these aspects. We were also intrigued by the prospect of working on a truly high-grade testbed environment that could allow us to complete a large number of experiments cycling through a potentially huge parameter space to provide meaningful results that can be applied in real-life situations. Having direct access to the tools and researchers involved with the use case also played a serious part in our motivation and we expect this can result in multiple mutual benefits.
How do you envision this project making an impact?
Our experiments on the testbed and with the tools of the SPIRIT project will provide insights for the scientific community about the effectiveness of QUIC used for a real-time telepresence application. We’ll release all our code as open-source facilitating further innovation.
Furthermore, our experiment is useful not just in academia, but it has industrial applications as well. We contribute to a scalable media server project, where one of the biggest challenges is the efficient ingesting of WebRTC traffic into a Kubernetes cluster. The results are going to serve as a stepping stone to replace the legacy WebRTC with QUIC-based video, leading to a scalable architecture for telepresence applications.
The SPIRIT project can also benefit from simply switching to QUIC as, for example, it allows for seamless mobility. Hence, it is also an enabler for new telepresence applications involving mobile collaboration.