Toward Adaptive and High-Performance XR Services via Network Programmability and Monitoring

Contemporary eXtended Reality (XR) services typically encounter performance and reliability limitations, especially when deployed over distributed environments with varying resources. This paper presents a modular and flexible XR-aware framework that (i) exposes network programmability interfaces to control different resources and segments in an unified manner, being capable of (ii) comprehensively monitoring and alerting cross-layer metrics from the XR service, network and cloud continuum infrastructure domains; and (iii) exploiting this comprehensive awareness to trigger dynamic (re-)configurations and mitigation actions, based on a set of criteria and Levels of Service (LoS) rules, with the goals of maximising the adaptivity, reliability and efficiency of the underlying XR services. Validation tests under a real Beyond 5G (B5G) testbed, derived from relevant use cases within the holographic communications domain, prove the efficiency and responsiveness of the proposed framework and anticipate its potential benefits.