Video Streaming Transport: Measurements and Advances

Saba Ahsan

Research output: ThesisDoctoral ThesisCollection of Articles

Abstract

Video streaming data forms the largest portion of the current Internet and is expected to rise even further. Providing and maintaining a high-quality experience for video is in the interest of both consumers and providers. However, the nature of network requirements for video traffic is different from other types of web traffic, and hence the goal requires a constant effort to advance video protocols in line with the changing Internet. This consists of not only developing new protocols and techniques to optimise the use of the Internet for video traffic but also developing meaningful video measurements that reflect user experience. In this thesis, we propose extensions and mechanisms to existing Internet protocols that enhance video quality and experience. We present results from experimentation that show how slight modifications to current protocols can be done to significantly improve the video streaming experience: (1) by introducing multiple path awareness (e.g. 3G, WiFi, Ethernet) to RTP, and (2) by using out of order delivery over TCP to improve timeliness for adaptive streaming. We also present a framework for enhancing current edge network measurements to encompass video experience using active testing while keeping the tests computationally light and easily deployable for large-scale network measurements. This thesis contributes to an evolving Internet, where video traffic has gained an equal if not higher importance to other web traffic. The large-scale video tests provide a mechanism where subscribers and Internet providers can better gauge performance from a video perspective, whereas, the work on adaptive video streaming over partially reliable and out of order transport paves the way for newer Internet protocols, which are now widely being recognised as a need of the hour in the heavily ossified and somewhat inflexible Internet.
Translated title of the contributionVideo Streaming Transport: Measurements and Advances
Original languageEnglish
QualificationDoctor's degree
Awarding Institution
  • Aalto University
Supervisors/Advisors
  • Ott, Jörg, Supervisor
Publisher
Print ISBNs978-952-60-8927-0
Electronic ISBNs978-952-60-8928-7
Publication statusPublished - 2020
MoE publication typeG5 Doctoral dissertation (article)

Keywords

  • multipath RTP
  • adaptive video streaming
  • DASH

Fingerprint Dive into the research topics of 'Video Streaming Transport: Measurements and Advances'. Together they form a unique fingerprint.

Cite this