Shaping the Future of XR: Innovations in Multi-Domain Orchestration

2023-11-08

In our modern interconnected world, where Augmented Reality (AR) and Virtual Reality (VR) technologies are anticipated to play a significant role, users expect seamless experiences across diverse devices and platforms. Nevertheless, ensuring those seamless experiences can pose a considerable challenge, particularly in distributed environments. To address these challenges, we are working on a new architectural multi-domain orchestration paradigm combining edge cloud technologies with AI solutions. 

This post highlights our take on meeting these expectations and overcoming the complexities of modern XR service delivery. 

The core of this architectural paradigm revolves around multi-domain edge-to-cloud orchestration, with a primary emphasis on Cloud-Native environments, Cloud-Native XR services, AI-driven deployment and lifecycle management, and a set of open-source enablers. This paradigm draws inspiration from the foundational principles of Zero-Touch Service Management (ZSM) architecture but is tailored to optimize and fulfil the demands of the next generation of extended reality (XR) services.

At the infrastructure level, we harness open-source technologies like Liqo, which fosters seamless interconnection of Kubernetes clusters across different domains, facilitating resource sharing and workload distribution. We also leverage ClusterAPI to extend and facilitate the orchestration of Kubernetes clusters on top of heterogeneous infrastructures and providers. Together, they form the basis of having a multi-domain computing substratum to deploy Cloud-based XR applications and services.

The concept of such interconnected domains and multi-domain orchestration aims to remove barriers from single-domain (and vendor) architectures but, at the same time, hide their complexities from the application developers. Furthermore, our approach considers the deployment and lifecycle management of services and infrastructure as potential autonomous and intelligent tasks that need coordination. For that, AI-driven closed-loop control systems, such as the OODA (Observe, Orient, Decide, Act) loops, can play a vital role.

For instance, in a live AR/VR streaming scenario, we continuously monitor the network and infrastructure conditions and XR-specific metrics such as user engagement and device performance. Upon a change, such as an increase in the number of viewers or a fluctuation in network quality, potentially leading to buffering or degradation of the user experience, OODA loops can swiftly take action to ensure a seamless and immersive AR/VR streaming experience by either decide to optimize video compression settings or switch to alternative content delivery routes and then.

This real-time AI-driven decision-making and resource orchestration unlocks the ultimate idea of a high-quality and uninterrupted AR/VR experiences, even when faced with unexpected network fluctuations or viewer numbers. 

We are shaping the future of XR service and infrastructure orchestration, we are at the forefront of innovation, 
OneSource and ICT-FI teams.