Innovative Algorithms for Prioritised AR/VR Content Delivery

Abstract

This paper proposes, describes and analyses two innovative approaches which make use of the Multipath Transmission Control Protocol (MPTCP) and its multiple flows for prioritised AR/VR content delivery. The first approach involves delivery of prioritised data using a fixed subflow and therefore establishment of a virtual private channel (VPC) for sending data. The second approach introduces a novel QoS on-the-fly (QoSF) algorithm that evaluates all subflows' delivery performance and dynamically selects one of them to deliver the prioritised content with the lowest latency. Both algorithms are assessed in singlehomed and multi-homed configurations and present different performance improvements in comparison to the classic MPTCP. While QoSF algorithm demonstrates the best performance in scenarios where the delivery latency variation between subflows is high, the VPC algorithm has the best results in scenarios where this variation is smaller.

A MPTCP-based RTT-aware Packet Delivery Prioritisation Algorithm in AR/VR Scenarios

“…RDPA improves a previous work in which a Quality of Service algorithm (QoSF) was introduced with a similar RTTaware behaviour [15]. RDPA introduces two new aspects to enhance the RTT-aware performance analysis.…”

“…RDPA monitors, identifies and performs packet delivery using the MPTCP subflow that can offer the best RTT (latency) for the prioritised content. RDPA enhances a previous work on a Quality of Service "On-the-fly" algorithm (QoSF) which also focuses on RTT-based data delivery [15]. RDPA introduces two new aspects to improve the RTT-aware performance analysis: validation of each subflow TCP window availability (subflow eligibility) and monitoring of each subflow average RTT value (subflow performance).…”

An Innovative Algorithm for Improved Quality Multipath Delivery of Virtual Reality Content

“…1) Linear regression: For the sake of a light-weight algorithm, this work extends a previous research [22], [27] and simplifies the linear regression applied to PHB report to extract only the slope of the linear regression for any given subflow feature performance (delay, throughout, etc.). Equation 1 describes how the slope is obtained (i.e., in this case, it is the covariance of x and y divided by the sum of squares (variance) of x).…”

Performance Evaluation of a Multi-User Virtual Reality Platform

“…Multi-user VR applications are designed for more than one user to be present in the virtual space at the same time. There are various parameters to be considered when developing such an application, as multi-user applications introduce challenges related to network performance, synchronisation among users and security to the system [6], [7], [8], [9].…”