Performance Analysis of Containerized Microservice Architectures for 360° Video Web Delivery Within an Edge-Cloud Infrastructure

Immersive content provides engaging experiences to users, with 360° videos being a prominent example. Such videos allow users to remotely explore events as if they were physically present. However, delivering these experiences incurs high computational costs. This challenge can be alleviated by distributing processing and communication tasks across an edge-cloud infrastructure. When combined with heterogeneous hardware and microservice architectures, this approach enables a scalable system for streaming 360° videos. In this paper we introduce a modular 360° video web streaming system deployed with various microservice architectures on an edge-cloud infrastructure. Our analysis reveals that edge-based architectures can cope with small-scale deployments, before hitting performance bottlenecks, by improving both delay and resource consumption. In contrast, when the video platform is stressed by adding more users, edge-cloud and microservice-assisted deployments become more efficient, improving both CPU and memory utilization, while providing similar delays as with a pure edge-based deployment.