Summary

Linux kernel statistics from a video server and service metrics from a video client

License

ODbL

Dependencies

Tags

audio-buffer-rate Linux-kernel network-analytics RTP-packet-rate service-level-metrics testbed time-series video-frame-rate video-on-demand VLC

Attribute Types

Download

gz (213.3 MB)

Files are converted on demand and the process can take up to a minute. Please wait until download begins.

Original Data Format

gz

Name

Version mldata

Comment

Names

Data (first 10 data points)

(Zipped) TAR archive

./._README.rtf, README.rtf, X_PROC, X_PROC/constant_load, X_PROC/flashcrowd_load, X_PROC/linear_increase, X_PROC/periodic_load, X_PROC/poisson_load, X_PROC/poisson_load/X.csv, X_PROC/poisson_load/Y.csv, X_PROC/periodic_load/X.csv, X_PROC/periodic_load/Y.csv, X_PROC/linear_increase/X.csv, X_PROC/linear_increase/Y.csv, X_PROC/flashcrowd_load/X.csv, X_PROC/flashcrowd_load/Y.csv, X_PROC/constant_load/X.csv, X_PROC/constant_load/Y.csv, X_SAR, X_SAR/constant_load, X_SAR/flashcrowd_load, X_SAR/linear_increase, X_SAR/periodic_load, X_SAR/poisson_load, X_SAR/poisson_load/X.csv, X_SAR/poisson_load/Y.csv, X_SAR/periodic_load/X.csv, X_SAR/periodic_load/Y.csv, X_SAR/linear_increase/X.csv, X_SAR/linear_increase/Y.csv, X_SAR/flashcrowd_load/X.csv, X_SAR/flashcrowd_load/Y.csv, X_SAR/constant_load/X.csv, X_SAR/constant_load/Y.csv

Description

Summary: Linux kernel statistics from a video server and service metrics from a video client

Tags: VLC Linux-kernel service-level-metrics network-analytics video-on-demand video-frame-rate audio-buffer-rate RTP-packet-rate time-series testbed

Description: Please cite this dataset as follows:

Yanggratoke, R., Ahmed, J., Ardelius, J., Flinta, C., Johnsson, A., Gillblad, D., & Stadler, R. (2014). Linux kernel statistics from a video server and service metrics from a video client. Distributed by Machine learning data set repository [MLData.org]. http://mldata.org/repository/data/viewslug/realm-im2015-vod-traces

Details description of the dataset including measurement details is available in the following publication. Yanggratoke, Rerngvit; Ahmed, Jawwad; Ardelius, John; Flinta, Christofer; Johnsson, Andreas; Gillblad, Daniel; Stadler, Rolf, "Predicting real-time service-level metrics from device statistics," Integrated Network Management (IM), 2015 IFIP/IEEE International Symposium on , vol., no., pp.414,422, 11-15 May 2015 doi: 10.1109/INM.2015.7140318

Field descriptions - Field descriptions for X.csv [1] "TimeStamp" : Unix epoch time [2..] X_SAR: http://linux.die.net/man/1/sar [2..] X_PROC: http://www.tldp.org/LDP/Linux-Filesystem-Hierarchy/html/proc.html

• Field descriptions for Y.csv [1] "TimeStamp" : Unix epoch time
  [2] "LostFrames" : Lost video frame rate (frames/sec)
  [3] “noAudioPlayed" : Audio buffer rate (buffers/sec)
  [4] "noAudioPlayedAvg2" : Average of [3] over a sliding window of 2 seconds (buffers/sec)
  [5] “noAudioPlayedAvg5" : Average of [3] over a sliding window of 5 seconds (buffers/sec)
  [6] “avgInterAudioPlayedDelay" : Average inter-audio-play delay (seconds) [7] "noAudioLost" : Lost audio buffer rate (buffers/sec)
  [8] “noAudioLate" : Late audio buffer rate (buffers/sec)
  [9] “NoRTPPkts" : RTP packet rate (packets/sec)
  [10] "AvgRTPInterPktDelay" : Average RTP inter packet delay (seconds) [11] “LostRTPPkts" : Lost RTP packet rate (packets/sec)
  [12] “AvgRTPJitter" : Average RTP jitter [13] "DispFrames" : Video frame rate (frames/sec) [14] “DispFramesAvg2” : Average of [13] over a sliding window of 2 seconds (buffers/sec)
  [15] “DispFramesAvg5" : Average of [13] over a sliding window of 5 seconds (buffers/sec)
  [16] "AvgInterDispDelay" : Average video-frame-inter-display delay (seconds)

URLs

http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-158063

Publications

• While real-time service assurance is critical for emerging telecom cloud services, understanding and predicting performance metrics for such services is hard. In this paper, we pursue an approach based upon statistical learning whereby the behavior of the target system is learned from observations. We use methods that learn from device statistics and predict metrics for services running on these devices. Specifically, we collect statistics from a Linux kernel of a server machine and predict client-side metrics for a video-streaming service (VLC). The fact that we collect thousands of kernel variables, while omitting service instrumentation, makes our approach service-independent and unique. While our current lab configuration is simple, our results, gained through extensive experimentation, prove the feasibility of accurately predicting client-side metrics, such as video frame rates and RTP packet rates, often within 10-15% error (NMAE), also under high computational load and across traces from different scenarios.

Data Source

Measurement Details

Usage Scenario

revision 1

by rerngvit on 2014-12-10 17:44

revision 2

by rerngvit on 2015-07-06 13:56

revision 3

by rerngvit on 2015-07-06 13:58

revision 4

by rerngvit on 2015-07-06 14:00

revision 5

by rerngvit on 2015-07-10 14:01

No one has posted any comments yet. Perhaps you would like to be the first?

Leave a comment

This item was downloaded 3258 times and viewed 3456 times.

No Tasks yet on dataset realm-im2015-vod-traces