Scalability Improvement Of Multicast Source Movement Over Mobile Ipv6 Using Clustering Technique

Mobile IPv6 (MIPv6) describes how a mobile node can change its point of attachment to the Internet. While MIPv6 focuses on unicast communications, it also proposes two basic mechanisms, known as bi-directional tunnelling and remote subscription, to handle multicast communications with mobile members. In the mean time, the deployment of Source-Specific Multicast (SSM) is of great interest, using the Protocol Independent Multicast-Sparse Mode (PIM-SM) and Multicast Listener Discovery (MLDv2) protocols. In the particular case of mobile IPv6 SSM sources, the mechanism proposed in MIPv6 to support multicast communications introduced a number of problems that need to be addressed. First, in most scenarios the MIPv6 solution leads to suboptimal routing by setting up a tunnel to forward packets between the home agent in its home network and the current location in the foreign network. The use of a third party when roaming which is the home agent leads to suboptimal routing. Second, it introduces a central point of failure (i.e. the Home Agent (HA)) that is not to be neglected. The proposed MIPv6 solution also induces a great traffic concentration around this central point. Third, the processing task of the central point increases with the number of mobile sources it serves, thus reducing the efficiency of multicast delivery. The objective of this thesis is to remove some of the obstacles encountered in the way of multicast deployment in the Internet, thereby making Mobile IPv6 better equipped to support mobile SSM sources. Recent proposals to provide multicasting over mobile IP focuses mainly on recipient mobility but little attention has been given to the case of source mobility. This thesis attempts to address this problem. The basic essence of the problem is that while the effect of receiver movement on the multicast tree is local, the effect of source movement may be global and it may affect the complete multicast delivery tree. The initial design was motivated by the need to support one-to-many and many-to-many applications in a scalable fashion. Such applications cannot be serviced efficiently with unicast delivery. As the overall problem statement of “Scalability Improvement of Multicast Source Movement over IPv6 Using Clustering Technique” is extremely complex, we divide the problem into the following components: build the multicast delivery tree for source specific multicast which is a routing issue; clustering receivers based on their IPv6 addresses; improve the state scalability of these clusters which is a deployment issue; find an efficient way for service distribution which is a deployment issue as well; and finally, the seamless integration of the work with Mobile IPv6 allowing it to support multicast efficiently for mobile nodes. The combined solution provides a comprehensive procedure for planning and managing a multicast-based IPv6 network. The outcome of this thesis are: a software to represent an architecture of a multicast delivery tree for one-to-many type of group communication, a group management scheme that could handle the end nodes subscription/un-subscription process with the required updates, an average subscription delay of between 0.255 ms-0.530 ms and un-subscription delay of between 0.0456 ms-0.087 ms for up to 50000 nodes, an approach to multicast forwarding state reduction that could support small-size groups as well as large-size groups, and finally the integration of the work with Mobile IPv6 to handle the multicast source movement.

Preview PDF FK_2006_106.pdf Download (565kB)

Actions (login required)

View Item