Content Delivery Infrastructures for Live Streaming Progress Report

Outline
Goals
Achievements
Forthcoming features
Participation

OpenCDN is an application-level Content Delivery Network for replication and splitting of live multimedia content and distributed caching of VoD, by deployment of a Streaming Relay Overlay Network. It has grown as an OpenSource project, and can be adapted to any kind of streaming technology. Free participation can be in the form of development and testing, set up of remote replication nodes, and contribution of live feeds.

When and if the OpenCDN project will be widely endorsed, netcasting to large audiences will become as easy as meeting friends at pub. No ubiquitous multicast routing will be necessary, and TV will become obsolete.

Artistic and cultural events will reach broader audiences, students will attend to courses from home, videoconferences will become Internet talk-shows.

Started in 2004, development still proceed unfounded (but I'm a bad manager). Despite this, two streaming technologies are actually supported, Apple Darwin Streaming Server, and Real Helix Universal Server, which can be hosted on the same machine. Contributing sources announce Metadata about their contents, and these are published at a Portal, where viewers can ask for an appropriate surrogate server. Request Routing, Distribution, and Delivery, is performed in the background by an XML-RPC based distributed architecture, which benefits of an inter-entity authentication infrastructure, and of a UDP probing mechanism, which also provides resilience against nodes and network outages.

Actual determination of a topologically efficient surrogate for a given viewer, is based on manually configured routing information. It will be substituted by an RTT estimate in between the candidate Last Hop Relays, and the viewer location

A web-based configuration interface will be developed in order to ease administration tasks

Better documentation will be provided, as well as How-To guides, about replication node and content source provision

Embedding of the requested content playback within the request page at the portal will be analyzed and developed

Aging of stale information will be performed, both at the centralized control entity, and at the distributed nodes

A policy agreement to be subscribed by contributors will be drawn, together with a community registration and communication paradigm

A graphical representation of the resulting topology will be provided, together with an easy access to statistical information about nodes usage

Installation and configuration of a replication node requires no more than a few hours work, and costs really nothing, as a cheap LinuxBox will suffice. Contents contribution just requires an ADSL connection, and only needs a publicly available multimedia encoder (such as VideoLAN, MPEG4IP, Real Producer). Development is made in Perl, and a rich ToDo list is present in the distribution. Code can be downloaded from SourceForge, where some mailing lists can be found also.