Split Rendering via SDP Signaling for Edge Cloud Offloading
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies face challenges in providing scalable and efficient rendering solutions for immersive media services like Metaverse, Virtual Reality, and Augmented Reality, particularly due to the high resource requirements for real-time rendering, which can overwhelm end devices and lead to increased latency and resource utilization.
Innovation Solution
The implementation of split rendering technology that offloads tasks from end devices to edge or cloud resources, utilizing Session Description Protocol (SDP) for optimized signaling between end devices, edge, and cloud to manage task distribution and resource allocation, enabling dynamic or static task splitting based on available resources and latency tolerance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If rendering tasks are processed locally on end devices, then real-time rendering performance is improved, but device resource requirements and complexity increase
Solution Approach 1:
The rendering task is divided into multiple independent parts (e.g., different rendering targets, scenes, or processing stages) that can be distributed across different devices. The system identifies which parts can be processed locally for real-time performance and which parts should be offloaded to remote resources, creating a segmented processing architecture that resolves the contradiction between speed and device complexity.
Solution Approach 2:
A cloud-based rendering service acts as an intermediary between the end device and the rendering workload. The intermediary receives rendering requests from end devices, processes them using remote computational resources, and returns results to the appropriate devices. This mediator enables high-performance rendering without requiring end devices to have the necessary computational resources.
2Device complexity
If rendering tasks are offloaded to cloud resources, then device resource requirements are reduced, but latency increases
Solution Approach 1:
Different parts of the rendering system have different quality requirements. Critical real-time rendering components are processed locally on end devices to maintain low latency, while computationally intensive but less time-sensitive tasks are offloaded to cloud resources. This local quality differentiation allows the system to optimize for both low latency and resource efficiency simultaneously.
Solution Approach 2:
The system performs preliminary processing and preparation of rendering tasks locally before offloading to cloud resources. By pre-processing scene graphs, preparing asset data, and establishing rendering contexts in advance, the system minimizes the actual transmission time and processing delay when tasks are sent to remote servers, thereby reducing overall latency despite the offloading operation.
3Ease of operation
If all rendering is done on end devices, then user experience is improved, but scalability of the service decreases
Solution Approach 1:
Instead of requiring full rendering capability on all end devices, the system implements partial rendering where only essential rendering functions are performed locally, while comprehensive rendering resources are provided by the cloud. This partial action approach maintains adequate user experience on consumer devices while enabling unlimited scalability through the cloud infrastructure that can accommodate any number of concurrent users.
Solution Approach 2:
The cloud-based rendering service provides universal rendering capability that can serve multiple types of devices and applications simultaneously. The same cloud infrastructure can support various rendering scenarios (VR, AR, 3D graphics, etc.) and different device types through standardized interfaces, enabling the service to scale universally across diverse use cases without requiring device-specific optimization for each function.
4Manufacturing precision
If high-resource devices are used for rendering, then rendering quality is improved, but device cost and accessibility decrease
Solution Approach 1:
Instead of providing expensive high-performance rendering hardware to each end device, the system creates virtual copies of rendering capabilities through cloud-based virtualization. Users access high-quality rendering services through software interfaces without needing the physical hardware, paying only for the computational resources consumed. This copying approach decouples rendering quality from device cost and enables universal access.
Solution Approach 2:
The system replaces the mechanical approach of equipping each device with powerful hardware accelerators with a software-based cloud rendering architecture. Instead of relying on local GPU hardware, the system uses virtualized rendering engines executed on remote servers, substituting physical hardware requirements with software service access. This substitution maintains high rendering quality while eliminating the need for expensive local hardware investments.
Data Source
AI summary
Facilitating elastic distributed computing for resource intensive tasks including split rendering in advanced networks is provided herein. Operations of a system include receiving information indicative of a request for split rendering for an application executing at a user equipment. The information is received via a session description protocol session. The operations can also include, based on the information indicative of the request, determining that an available computational resource capacity at the user equipment is insufficient to process the application executing at the user equipment. Further, the operations can include, based on the determining, enabling a first rendering of a first portion of the application at a first device and a second rendering of a second portion of the application at a second device different from the first device.