Abstract

The increasing variety of networks and end systems, especially wireless devices, pose new challenges in communication support for, particularly, multicast-based collaborative applications. In traditional multicasting, the sender transmits video at the same rate and resolution to all receivers independent of their network characteristics, end system equipment, and users' preferences about video quality and significance. Such an approach results in resources being wasted and may also result in some receivers having their quality expectations unsatisfied. This problem can be addressed, near the network edge, by applying dynamic, in-network adaptation (e.g., transcoding) of video streams to meet available connection bandwidth, machine characteristics, and client preferences. In this paper, we extrapolate from earlier work of Shorfuzzaman et al. 2006 in which we implemented and assessed an MPEG-1 transcoding system on the Intel IXP1200 network processor to consider the feasibility of in-network transcoding for other video formats and network processor architectures. The use of “on-the-fly” video adaptation near the edge of the network offers the promise of simpler support for a wide range of end devices with different display, and so forth, characteristics that can be used in different types of environments.

Highlights

  • The rapid growth of distributed computing and the Internet has led to demand for collaboration over wide area networks

  • This demand has been only partially met by existing multimedia and collaborative applications such as video-ondemand, teleconferencing, and telemedicine, which use the Internet for communication

  • Responding to this issue, the problem addressed in this paper is “how can network performance in collaborative applications be improved by detecting and managing preferences from the receivers for use in dynamic, in-network adaptation of data streams?” To this end, we discuss a framework that addresses network heterogeneity by considering clients’ network connections and device characteristics and supports delivering activity-based user interest hints into the network and using those hints at routers to adapt to changing user requirements through the dynamic modification of data streams

Read more

Summary

Introduction

The rapid growth of distributed computing and the Internet has led to demand for collaboration over wide area networks. Independent of their network attachment and end systems equipment This means the source can only generate data at a rate that meets the capability of the most constrained receiver, receivers having high bandwidth links would be capable of receiving correspondingly higher quality video streams. We seek to support the use of a wide range of end devices, varying connection characteristics and different user interests through the use of in-network video transcoding techniques implemented near the network edge in a fashion similar, but not limited, to that of active networks. Our current implementation uses the IXP1200 network processor for implementing the nodes that transcode MPEG-1 (MPEG-1 was chosen due to the limited instruction store and processing capability of the IXP1200.) video data to a desired bit rate.

Related Work
Different Video Coding Standards
Video Adaptation Algorithms
Frame number
Low-Pass Filter
Frame Resizing
Network Processors for Video Transcoding
Overview of Network Processors
Implementation of the Active Video Adaptation Node
10. Protoype Evaluation
Threads 281 214 338 231
Threads 1450 1620 1634 1470 1402 1650
Threads 1590 1713 1506 1669
Findings
11. Conclusions
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call