Show simple item record

AuthorBilal K.
AuthorErbad A.
AuthorHefeeda M.
Available date2020-02-24T08:57:14Z
Publication Date2018
Publication NameJournal of Network and Computer Applications
ResourceScopus
ISSN10848045
URIhttp://dx.doi.org/10.1016/j.jnca.2018.07.012
URIhttp://hdl.handle.net/10576/13023
AbstractVideo streaming is one of the most prevailing and bandwidth consuming Internet applications today. Advancements in technology and prevalence of video capturing devices result in massive multi-sourced (aka crowdsourced) live video broadcasting over the Internet. A single scene may be captured by multiple spectators from different angles (views), enabling an opportunity for interactive multiview video by integrating these individually captured views. Such multi-sourced multiview video offers more realistic and immersive experience of a scene. In this paper, we present a Quality of Experience (QoE) driven, cost effective Crowdsourced Multiview Live Streaming (CMLS) system. The CMLS aims to minimize the overall system cost by selecting optimal cloud site for video transcoding and the number of representations, based on the view popularity and viewer's available bandwidth. In addition, we present a QoE metric considering delay and received video quality. We formulate the selection of optimal cloud site and number of representations to meet the required QoE as a resource allocation problem using Integer Programming (IP). Moreover, we present a Greedy Minimal Cost (GMC) algorithm to perform resource allocation efficiently. We use real live video traces collected from three large-scale live video providers (Twitch.tv, YouTube Live, and YouNow) to evaluate our proposed strategy. We evaluate the GMC algorithm considering the overall cost, QoE, video quality, and average latency between viewers and transcoding location. We compare our results with the optimal solution and the state-of-the art policy used in a popular video steaming system. Our results demonstrate that the GMC achieves near optimal results and substantially outperforms the state-of-the art policy.
SponsorThis publication was made possible by NPRP grant # [ 8-519-1-108 ] from the Qatar National Research Fund (a member of Qatar Foundation). We are thankful to the Denny Stohr for providing YouNow dataset. The findings achieved herein are solely the responsibility of the author[s].
Languageen
PublisherAcademic Press
SubjectCloud computing
Crowdsourced live video
Multiview video streaming
Resource allocation
TitleQoE-aware distributed cloud-based live streaming of multisourced multiview videos
TypeArticle
Pagination130 - 144
Volume Number120
dc.accessType Abstract Only


Files in this item

FilesSizeFormatView

There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record