MAGNOLIA

Bringing an immersive experience to the Internet

Immersive digital applications, such as 360° video, extended reality (XR), remote training, and virtual conferencing, place high demands on networks and systems. They require, among others, a very low delay (latency). Even small delays of a few tens of milliseconds can hamper the user experience and cause motion sickness, reduce realism, or make the interaction difficult.

Today, even with 5G technology, most immersive applications only work well on local networks that are tightly controlled, fast, and lightly loaded. In contrast, on the open Internet traffic is shared by many applications, data flows through several network segments, and congestion can occur at many points. As a result, latency builds up quickly and unpredictably.

The MAGNOLIA project aims to address these challenges with a focus on end-to-end low latency. The goal is to enable future 6G networks to run real-time immersive applications reliably over the open Internet. MAGNOLIA wants to accomplish this through three tightly connected innovation goals.

MAGNOLIA’s three innovation goals

First, we want to achieve end-to-end low latency over an open network. To do so, we will develop a framework for next-generation radio access networks (RAN) that is Internet Protocol (IP)-aware. RANs enable mobile communication between end-user apps and network base stations. Making them IP-aware allows them to be conscious of application traffic and adapt accordingly. We expect that the 6G architecture will be flexible enough to support this.

Next, the idea is to reduce the delay inside the immersive applications themselves, improving how XR and 360° video are captured, encoded, sent, decoded, and rendered. The content quality will be adapted based on user focus and context, so that less important parts require fewer computing and network resources.

Third, we will ensure that applications, networks, and cloud work tightly together. Applications will use feedback from the network about congestion and available capacity and adapt their behaviour in real time. In addition, there will be dynamic decisions about where in the network processing should happen, allowing for the lowest latency and best efficiency. In particular, we want to implement multi-path traffic control, splitting traffic across paths based on real-time network feedback and guaranteeing high throughput and low-latency for time-critical traffic flows. In addition, machine learning will help manage complex traffic patterns.

Project outcomes

MAGNOLIA aims to deliver a framework for low-latency immersive applications over the open Internet. The project partners will validate their results through two real-world demonstrators.

The first demonstrator focuses on remote maintenance where users interact with XR content in real time over multiple wireless networks. This system demonstrates how low latency enables natural interaction, precise manipulation, and shared experiences at a distance.

The second demonstrator focuses on immersive conferencing and simultaneous language interpretation. 360° video is used to capture rich visual context. Intelligent viewpoint selection and framing ensure that the interpreters and other participants receive the information they need with minimal delay. That way, conference rooms can be connected over long distances while preserving natural communication. 

“MAGNOLIA enables real-time immersive applications over the open Internet by redesigning networks, applications, and cloud systems to work together with ultra-low latency, making technologies like XR, 360° video, and remote collaboration reliable beyond local networks.”