On the Load Balancing of Edge Computing resources for on-line video delivery

Utku Bulkan, Tasos Dagiuklas, Muddesar Iqbal, Kazi Mohammed Saidul Huq, Anwer Al-Dulaimi, Jonathan Rodriguez

Research output: Contribution to journalArticlepeer-review

8 Citations (Scopus)

Abstract

Online video broadcasting platforms are distributed, complex, cloud oriented, scalable, micro-service based systems that are intended to provide Over-The-Top (OTT) and live content to audience in scattered geographic locations. Due to the nature of cloud VM hosting costs, the subscribers are usually served under limited resources in order to minimize delivery budget. However, operations including transcoding require high computational capacity and any disturbance in supplying requested demand might result in Quality of Experience (QoE) deterioration. For any online delivery deployment, understanding users QoE plays a crucial role for rebalancing cloud resources. In this work, a methodology for estimating Quality of Experience is provided for a scalable cloud based online video platform. The model will provide an adeptness guideline regarding limited cloud resources and relate computational capacity, memory, transcoding and throughput capability and finally latency competence of the cloud service to QoE. Scalability and efficiency of the system are optimized through reckoning sufficient number of VMs and containers to satisfy the user requests even on peak demand durations with minimum number of VMs. Both horizontal and vertical scaling strategies (including VM migration) are modelled to cover up availability and reliability of intermediate and edge Content Delivery Network (CDN) cache nodes
Original languageEnglish
Article number8554079
Pages (from-to)73916-73927
Number of pages12
JournalIEEE Access
Volume6
DOIs
Publication statusPublished - 30 Nov 2018

Bibliographical note

Publisher Copyright:
© 2018 IEEE.

Keywords

  • Availability
  • Cloud
  • Content management systems
  • Dockers
  • Mathematical modeling
  • Online video platform
  • QoE
  • Reliability
  • Scalability
  • Virtual machines

Fingerprint

Dive into the research topics of 'On the Load Balancing of Edge Computing resources for on-line video delivery'. Together they form a unique fingerprint.

Cite this