Low Latency Low Loss Media Delivery Utilizing In-Network Packet Wash


연구 분야: Networking



학회: Journal of Network and Systems Management


초록

This paper presents new techniques and mechanisms for carrying streams of layered video using Scalable Video Coding (SVC) from servers to clients, utilizing the Packet Wash mechanism which is part of the Big Packet Protocol (BPP). BPP was designed to handle the transfer of packets for high-bandwidth, low-latency applications, aiming to overcome a number of issues current networks have with high precision services. One of the most important advantages of BPP is that it allows the dynamic adaption of packets during transmission. BPP uses Packet Wash to reduce the payload, and the size of a packet by eliminating specific chunks. For video, this means cutting out specific segments of the transferred video, rather than dropping packets, as happens with UDP based transmission, or retrying the transmission of packets, as happens with TCP. The chunk elimination approach is well matched with SVC video, and these techniques and mechanisms are utilized and presented. An evaluation of the performance is provided, plus a comparison of using UDP or TCP, which are the other common approaches for carrying media over IP. Our main contributions are the mapping of SVC video into BPP packets to provide low latency, low loss delivery, which provides better QoE performance than either UDP or TCP, when using those techniques and mechanisms. This approach has proved to be an effective way to enhance the performance of video streaming applications, by obtaining continuous delivery, while maintaining guaranteed quality at the receiver. In this work we have successfully used an H264 SVC encoded video for layered video transmission utilizing BPP, and can demonstrate video delivery with low latency and low loss in limited bandwidth environments.


Author Profile
Stuart Clayman

Department of Electronic and Electrical Engineering University College London London UK

Andorra
Author Profile
Müge Sayıt

International Computer Institute Ege University Izmir Turkey

Turkey

📄 논문 정보

발행 연도 2023년
인용수 12
출판 국가 Andorra, Turkey
사이트 Springer
좋아요 수 0

연관 논문 목록 (19건)