An Architecture for Seamless Access to IP Multicast Content

Despite the tremendous success of the Internet, we are still using quite inefficient methods to distribute Internet content to a large number of simultaneous receivers. This Internet broadcasting, does not scale when one uses the traditional Internet model of point-to-point unicast communication. As a result, the networking research community proposed in the late 80’s IP multicast. IP multicast is a network layer service that allows a single source to distribute a data stream to many simultaneous receivers in an efficient manner. However, this technology is not widely implemented in the Internet, mainly due to the prohibitively high deployment and management cost of IP multicast. This high cost has for effect that an ISP already needs a potentially high number of IP multicast capable end-users before IP multicast becomes more profitable than using unicast for serving these users. As a result, in spite of more than a decade of existence, the IP multicast architecture remains just a research product with very limited penetration into the commercially deployed Internet.In this dissertation, we propose partially transitional additions to the existing IP multicast service model, which transforms IP multicast in to a technology which can then become customer pulled. The architecture is composed of two sub parts: the first involves the access to IP multicast content and is called CastGate; the second part is CastGuide, which offers the technology to build an enhanced IP multicast “program guide”. The CastGate technology gives Internet users access to IP multicast content even when connected to a unicast-only ISP. It is built around the concept of application level tunneling augmented with an automatic tunnel end-point location mechanism. This solution can be seen as a transition technology, which artificially creates multicast-capable end-users in order to smoothen the IP multicast deployment process. At the same time it gave us the possibility to address issues lacking in the current IP multicast Service Model. We added support for user authentication, authorization and accounting and we also fitted the architecture with a Congestion Control tunnel framework to guarantee TCP-friendliness of the tunnel flows. The CastGate architecture requires no modification at the level of the network or at the end-systems. It does not necessarily require cooperation from the ISP and its deployment is as simple as installing web servers. The client part can be a simple (signed) JAVA applet running in the users’ web browser or a separate tool with integrated tunnel support. The technology allows the end-user to receive and source “traditional” IP multicast traffic (ASM) and Single Source Multicast (SSM) style of content. It is as good as the “real thing” from the end-users’ perspective.At the same time we also investigated more general problems with which end-users are faced when accessing IP multicast Content, either natively or via a CastGate tunnel. As a result we created a new generation of Session Directory tool called CastGuide with integrated CastGate client technology. This IP multicast “program guide” uses SDP Proxies to quickly present end-users with up-to-date lists of available and upcoming IP multicast sessions. Additionally we looked at issues faced by Content Providers when announcing their sessions. In response we propose the addition of the “Channel” directory type to ensure that a Content Provider is granted exclusive rights on the announcements sub-tree. We are convinced that the introduction of the CastGuide and CastGate technologies can make it for end-users more attractive to access IP multicast Content and make it at the same time