Online Multicast Traffic Engineering for Multi-View Videos With View Synthesis in SDN

Multi-view videos (MVV) have emerged to provide users with immersively interactive experiences with 3D multimedia content. Compared with traditional 2D videos, MVV offers multiple view angles to avoid generating occluded regions from a single viewpoint and allows users to receive different view angl...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:IEEE/ACM transactions on networking Ročník 32; číslo 4; s. 2778 - 2793
Hlavní autoři: Chiang, Sheng-Hao, Wang, Chih-Hang, Yang, De-Nian, Liao, Wanjiun, Chen, Wen-Tsuen
Médium: Journal Article
Jazyk:angličtina
Vydáno: IEEE 01.08.2024
Témata:
Bandwidth
Costs
Electronic mail
multi-view video
multicast
Multicast algorithms
online algorithm
SDN
Servers
Urban areas
Videos
ISSN:1063-6692, 1558-2566
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Multi-view videos (MVV) have emerged to provide users with immersively interactive experiences with 3D multimedia content. Compared with traditional 2D videos, MVV offers multiple view angles to avoid generating occluded regions from a single viewpoint and allows users to receive different view angles, which consume much higher bandwidth. In this paper, we leverage multicast and view synthesis to effectively reduce the number of transmitted views and total bandwidth consumption in software-defined content delivery networks (SD-CDN). By exploiting the SDN architecture, SD-CDN can optimize traffic engineering and the selection of multi-view sources to serve users. We formulate a new optimization problem, Online Multicast with View Synthesis (OMVS), prove the NP-hardness, and design an online algorithm with the ideas of View Popularity Cost Ratio, View Watching Possibility, and synthesis tree, to achieve the tightest competitive ratio. The experiment on real networks and implementation in an experimental SDN manifest that the proposed algorithm outperforms state-of-the-art algorithms regarding the total cost, bandwidth consumption, synthesis quality, and link and node utilization.
ISSN:1063-6692
1558-2566
DOI:10.1109/TNET.2024.3366166