An Extensive Distributed Measurement Platform for Analyzing Quality-of-Experience (QoE) of Multimedia Applications over Wireless Networks


The primary objective of this project is to develop a software architecture for distributed measurement system that analyzing Quality of Experience (QoE) of multimedia applications over wireless network and monitor network Quality of Service (QoS). Aiming to be implemented on a mobile handset across different Smartphone platforms, the application system serves as a convenient mobile network monitoring tool for both end-users and network operators. Time periods and cells that are likely to have network congestion can be located, and provide valuable information for improving satisfaction level of service subscribers.


Overall Architecture

The architecture consists of mobile and PC components and a set of APIs. There are two components in our final product, namely (1) mobile phone application; and (2) a server providing backend support. The mobile phone application comprises four major software components, namely (1) Measurement Platform; (2) Network Monitoring Tool; (3)Quality of Service (QoS) module; and (4)Quality of Experience (QoE) module.


Measurement Platform

The measurement platform serves as a central hub for network performance data collection. It has a built-in abstraction layer for interfacing with Smartphone hardware across platforms. Radio information like serving Cell-ID, radio signal strength, and location information (though GPS or network-assisted GPS) are monitored continuously. These data are recorded along with the network statistics and time in the data logs, which will be important for tracing network congestion zones. The measurement platform may also invoke its two sub-modules, namely QoE module and QoS module, to assess QoE and QoS information. Furthermore, the measurement platform uploads the collected data to the backend server via Data Upload API for long-term storage. These data uploads are carried out mainly during idle periods to minimize performance impact to normal phone usage. To support network measurement at idle periods or at intervals triggered by specific events (e.g. radio signal strength below a certain threshold, handset under coverage of a particular cell), a rule engine is incorporated in the measurement platform. The rule-engine from time to time evaluates its rule inventory (defined by Sensor Rules API) and triggers network statistics measurement if needed. As a result, dynamic update to the rules for triggering measurement may modify the measurement schedule. For instance, a simple modification of the rule-set (e.g. by SMS message, via SMS API) from the network operator triggers a server-initiated measurement.

Network Monitoring Tool

The network monitoring tool provides access to information related to current network performance, receives data from the measurement platform, and collects basic phone status readings, including QoS and QoE measures. Its graphical interface allows Smartphone users or network operator engineers to manually trigger network performance measurement. Periodic calls are sent to the measurement platform via MP API to maintain a list of updated network performance metrics. The tool also interacts with the backend server for functions like historical data retrieval for presentation to users. All the collected data are eventually uploaded to the backend server for persistent storage.

QoS Module and QoE Module

The QoS module, a sub-component of the measurement platform, collects various QoS metrics. include uplink / downlink throughput, round trip time (RTT), to be carried out during idle periods as well as during multimedia streaming sessions.

As another sub-component of measurement platform, the QoE module takes up QoE-relevant tasks. There will be two processing modes for QoE evaluation, online and offline, depending on the precision required and the corresponding performance impact on mobile devices. For online processing, a lightweight version of a perceptual QoE assessment scheme will be implemented on mobile devices, within the QoE module. This assessment tool will be used to evaluate in real-time, the perceived QoE of the received multimedia streams. For offline processing, packet arrival statistics are captured to be sent back to the backend server (mediated by the measurement platform) for offline QoE evaluation.

Backend Server

The backend server mainly serves two purposes. Firstly, it receives network performance data from the measurement platform for persistent storage. From time to time, it sends historical data back to the mobile phone upon request by the network monitoring tool. Secondly, with arrival packet data captured by the QoE module on mobile phone, the server reconstructs the received multimedia stream to evaluate the perceived QoE using different state-of-the-art QoE assessment schemes.

Smartphones Screen Captures

Symbian Screen Capture
Android Screen Capture

.: TOP :.
Last Updated on October 15 2015.
´╗┐Copyright ┬ę 2015. All Rights Reserved. MobiTeC, The Chinese University of Hong Kong.