Online architecture for predicting live video transcoding resources

Research output: Contribution to journalArticleScientificpeer-review

Abstract

End users stream video increasingly from live broadcasters (via YouTube Live, Twitch etc.). Adaptive live video streaming is realised by transcoding different representations of the original video content. Management of transcoding resources creates costs for the service provider, because transcoding is a CPU-intensive task. Additionally, the content must be transcoded within real time with the transcoding resources in order to provide satisfying Quality of Service. The contribution of this paper is validation of an online architecture for enabling live video transcoding with Docker in a Kubernetes-based cloud environment. Particularly, online cloud resource allocation has been focused on by executing experiments in several configurations. The results indicate that Random Forest regressor provided the best overall performance in terms of precision regarding transcoding speed and CPU consumption on resources, and the amount of realised transcoding tasks. Reinforcement Learning provided lower performance, and required more effort in terms of training.

Original languageEnglish
Article number9
JournalJournal of Cloud Computing
Volume8
Issue number1
DOIs
Publication statusPublished - 2019
MoE publication typeA1 Journal article-refereed

Fingerprint

Program processors
Video streaming
Reinforcement learning
Resource allocation
Quality of service
Costs
Experiments

Keywords

  • Cassandra
  • Docker
  • FFmpeg
  • Gym
  • Rancher
  • Random Forest
  • Reinforcement learning
  • RL-Keras

Cite this

@article{71f8f20a4490472bb129e5714ce7d0a5,
title = "Online architecture for predicting live video transcoding resources",
abstract = "End users stream video increasingly from live broadcasters (via YouTube Live, Twitch etc.). Adaptive live video streaming is realised by transcoding different representations of the original video content. Management of transcoding resources creates costs for the service provider, because transcoding is a CPU-intensive task. Additionally, the content must be transcoded within real time with the transcoding resources in order to provide satisfying Quality of Service. The contribution of this paper is validation of an online architecture for enabling live video transcoding with Docker in a Kubernetes-based cloud environment. Particularly, online cloud resource allocation has been focused on by executing experiments in several configurations. The results indicate that Random Forest regressor provided the best overall performance in terms of precision regarding transcoding speed and CPU consumption on resources, and the amount of realised transcoding tasks. Reinforcement Learning provided lower performance, and required more effort in terms of training.",
keywords = "Cassandra, Docker, FFmpeg, Gym, Rancher, Random Forest, Reinforcement learning, RL-Keras",
author = "Pekka P{\"a}{\"a}kk{\"o}nen and Antti Heikkinen and Tommi Aihkisalo",
year = "2019",
doi = "10.1186/s13677-019-0132-0",
language = "English",
volume = "8",
journal = "Journal of Cloud Computing",
issn = "2192-113X",
publisher = "Springer",
number = "1",

}

Online architecture for predicting live video transcoding resources. / Pääkkönen, Pekka; Heikkinen, Antti; Aihkisalo, Tommi.

In: Journal of Cloud Computing, Vol. 8, No. 1, 9, 2019.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - Online architecture for predicting live video transcoding resources

AU - Pääkkönen, Pekka

AU - Heikkinen, Antti

AU - Aihkisalo, Tommi

PY - 2019

Y1 - 2019

N2 - End users stream video increasingly from live broadcasters (via YouTube Live, Twitch etc.). Adaptive live video streaming is realised by transcoding different representations of the original video content. Management of transcoding resources creates costs for the service provider, because transcoding is a CPU-intensive task. Additionally, the content must be transcoded within real time with the transcoding resources in order to provide satisfying Quality of Service. The contribution of this paper is validation of an online architecture for enabling live video transcoding with Docker in a Kubernetes-based cloud environment. Particularly, online cloud resource allocation has been focused on by executing experiments in several configurations. The results indicate that Random Forest regressor provided the best overall performance in terms of precision regarding transcoding speed and CPU consumption on resources, and the amount of realised transcoding tasks. Reinforcement Learning provided lower performance, and required more effort in terms of training.

AB - End users stream video increasingly from live broadcasters (via YouTube Live, Twitch etc.). Adaptive live video streaming is realised by transcoding different representations of the original video content. Management of transcoding resources creates costs for the service provider, because transcoding is a CPU-intensive task. Additionally, the content must be transcoded within real time with the transcoding resources in order to provide satisfying Quality of Service. The contribution of this paper is validation of an online architecture for enabling live video transcoding with Docker in a Kubernetes-based cloud environment. Particularly, online cloud resource allocation has been focused on by executing experiments in several configurations. The results indicate that Random Forest regressor provided the best overall performance in terms of precision regarding transcoding speed and CPU consumption on resources, and the amount of realised transcoding tasks. Reinforcement Learning provided lower performance, and required more effort in terms of training.

KW - Cassandra

KW - Docker

KW - FFmpeg

KW - Gym

KW - Rancher

KW - Random Forest

KW - Reinforcement learning

KW - RL-Keras

UR - http://www.scopus.com/inward/record.url?scp=85068774844&partnerID=8YFLogxK

UR - https://rdcu.be/bJAgl

U2 - 10.1186/s13677-019-0132-0

DO - 10.1186/s13677-019-0132-0

M3 - Article

VL - 8

JO - Journal of Cloud Computing

JF - Journal of Cloud Computing

SN - 2192-113X

IS - 1

M1 - 9

ER -