QoS-QoE Dataset - Predicting QoE Factors with Machine Learning (IEEE ICC 2018)

This page presents the dataset used in our ICC 2018 paper:

Predicting QoE Factors with Machine Learning.
Vladislav Vasilev, Jeremie Leguay, Stefano Paris, Lorenzo Maggi, Merouane Debbah.
IEEE ICC 2018. Kansas City, USA. May 2018. pdf
Abstract: Classic network control techniques have as sole objective the fulfillment of Quality-of-Service (QoS) metrics, being quantitative and network-centric. Nowadays,the research community envisions a paradigm shift that will put the emphasis on Quality of Experience (QoE) metrics, which relate directly to the user satisfaction. Yet, assessing QoE from QoS measurements is a challenging task that powerful Software Defined Network controllers are now able to tackle via machine learning techniques. In this paper we focus on a few crucial QoE factors and we first propose a Bayesian Network (BN) model to predict re-buffering ratio. Then, we derive our own novel Neural Network search method to prove that the BN correctly captures the discovered stalling data patterns. Finally, we show that hidden variable models based and context information boost performance for all QoE related measures.

To build and evaluate QoS to QoE mapping functions, we have used a fully controllable simulation environment at both network and streaming levels. The simulation platform is based on the Adaptive Multimedia Streaming Simulator Framework (AMust) in NS3 which implements a HTTP client and server for LibDASH, one of the reference software of ISO/IEC MPEG-DASH standard.

Simulation parameters with AMust in NS3
On the star topology with a bottleneck link, as depicted in the figure above, we have simulated a large number of streaming scenarios varying the number of streaming clients and the bottleneck characteristics (capacity, delay, packet loss). After a month of simulations, we have obtained statistics for more than 69,129 video sessions with 51 features for each session. For each video session, we measure network and streaming level statistics:

Column	Name	Type	Description
1	RequestID	Context	Streaming session identifier
2	NbClients	Context	Maximum number of streams competing on the bottleneck
3	BottleneckBW	Context	Capacity of the bottleneck
4	BottleneckDelay	Context	Network delay on the bottleneck
5	BottleneckLoss	Context	Packet loss on the bottleneck
6	DASHPolicy	Context	DASH policy
7	ClientResolution	Context	Client screen resolution
8	RequestDuration	Context	Duration of the stream
[9,13]	TCPOut/InPacket	QoS metric	Number of TCP packets (In and Out)
[10,14]	TCPOut/InDelay	QoS metric	Average delay experienced by TCP packets (In and Out)
[11,15]	TCPOut/InJitter	QoS metric	Average jitter experienced by TCP packets (In and Out)
[12,16]	TCPOut/InPLoss	QoS metric	Packet loss rate experienced by TCP packets (In and Out)
17	TCPInputRetrans	QoS metric	Packet retransmissions experienced by TCP
18	StdNetworkRate	QoS metric	Standard deviation of the network rate
[19:27]	[0,5,10,25,50,75,90,95,100] NetworkRate	QoS metric	Quantiles for the network rate
28	StdInterATimesReq	QoS metric	Std. dev. of inter-arrival times of segment requests
[29:37]	[0,5,10,25,50,75,90,95,100] InterATimesReq	QoS metric	Quantiles for inter-arrival times of segment requests
38	StartUpDelay	QoE metric	Initial time at the client to start playing the video
39	AvgVideoDownloadRate	QoE metric	Average downloading rate for video segments
40	StdVideoDownloadRate	QoE metric	Std. dev. of downloading rate for video segments
41	AvgVideoBufferLevel	QoE metric	Average video buffer length
42	StdVideoBufferLevel	QoE metric	Std. dev. of video buffer length
43	StallEvents	QoE metric	Number of stall events
44	RebufferingRatio	QoE metric	Portion of time spent in stall events
45	StallLabel	QoE metric	Categorization of RebufferingRatio variable
46	TotalStallingTime	QoE metric	Total duration of stall events
47	AvgTimeStallingEvents	QoE metric	Average duration of stall events
48	AvgQualityIndex	QoE metric	Avg. normalized index of downloaded representations
49	AvgVideoBitRate	QoE metric	Average video bitrate consumed by the player
50	AvgVideoQualityVariation	QoE metric	Average variation of the video bitrate
51	AvgDownloadBitRate	QoE metric	Average download rate of video segments

The full dataset is available here: zip file

The simulation code is available here: https://github.com/sassatelli/QoErouting

When writing a paper that uses this Code, we would appreciate that you cite the following papers:

Predicting QoE Factors with Machine Learning.
Vladislav Vasilev, Jeremie Leguay, Stefano Paris, Lorenzo Maggi, Merouane Debbah.
IEEE ICC 2018. Kansas City, USA. May 2018. pdf
AMuSt Framework - Adaptive Multimedia Streaming Simulation Framework for ns-3 and ndnSIM.
Christian Kreuzberger, Daniel Posch, Hermann Hellwagner.
https://github.com/ChristianKreuzberger/AMust-Simulator/

The simulation environment has also been used in our QoE-aware routing project:

Quality of Experience-based Routing of Video Traffic for Overlay and ISP Networks.
Giacomo Calvigioni, Ramon Aparicio-Pardo, Lucile Sassatelli, Jeremie Leguay, Stefano Paris, Paolo Medagliani.
IEEE INFOCOM 2018. Hawaii, USA. April 2018. pdf