Fuzzycast: efficient video-on-demand over multicast

Abstract

Server bandwidth has been identified as a major bottleneck in large Video-on-Demand (VoD) systems. Using multicast delivery to serve popular content helps increase scalability by making efficient use of server bandwidth. In addition, recent research has focused on proactive schemes in which the server periodically multicasts popular content without explicit requests from clients. Proactive schemes are attractive because they consume bounded server bandwidth irrespective of client arrival rate.In this work, we describe Fuzzycast, a scalable periodic multicast scheme that uses simple techniques to provide video on demand at reasonable client start-up times while consuming optimal server bandwidth. We present a theoretical analysis of its bandwidth and client buffer requirements and prove its optimality. We study the effect of variable bitrate (VBR) media on Fuzzycast performance and propose a simple extension to transmit VBR media over constantrate channels. Finally, we solve the problem of partitioning a transmission over multiple multicast groups by considering it as a specific instance of a more widely encountered resource trade-off.

<title>Efficient Buffer Management for Scalable Media-on-Demand</title>

“…It can be proved [7] that the average bandwidth for the entire movie is: In practical terms, serving a 2-hour 300 kbps Real Media or MPEG-4 movie with a 5-minute initial delay requires a server and client bandwidth of ≈ 1 Mbps. Thus, the system begins to be advantageous as soon as the number of clients exceeds 3.…”

“…THE BUFFER MANAGEMENT PROBLEM As it stands, Algorithm IDEAL results in too spiky a bandwidth usage to be implemented in practice. We have developed schemes [7] that schedule frame transmission in a way that preserves its bandwidth optimality without violating peak bandwidth constraints. These schemes work by "fuzzifying" the schedules, redistributing bandwidth more equitably over time by moving frames away from bandwidth "spikes".…”

“…We do not discuss further the ways in which FIND-NEIGHBOUR may be implemented, but refer the interested reader instead to [7]. For this work , it is sufficient to assume that for a movie of n frames, any client with a join time of t s and an initial wait period of w would receive at least one instance of any given frame f at some instant t f before playout, i.e.…”

“…transmit (t + δ t , f j ); end for end for be shown [7] that the client buffer size as a function of time t relative to t s is…”

“…Earlier work [1] on requests in video rentals suggests that an 80:20 rule might hold here: 80% of requests are for the top 20 movies. Recent research [2], [3], [4], [5], [6], [7] has thrown up a variety of scalable MoD schemes that attempt to conserve server bandwidth by segmenting popular MoD data and periodically broadcasting it to interested clients.…”