Developing the Fringe Routing Protocol

<p>An ISP style network often has a particular traffic pattern not typically seen in other networks and which is a direct result of the ISP’s purpose, to connect internal clients with a high speed external link. Such a network is likely to consist of a backbone with the clients on one ‘side’ and one or more external links on the other. Most traffic on the network moves between an internal client and the external world via the backbone. But what about traffic between two clients of the ISP? Typical routing protocols will find the ‘best’ path between the two gateway routers at the edge of the client stub networks. As these routers connect the stubs to the ISP core, this route should be entirely within the ISP network. Ideally, from the ISP point of view, this traffic will go up to the backbone and down again but it is possible that it may find another route along a redundant backup path. Don Stokes of Knossos Networks has developed a protocol to sit on the client fringes of this ISP style of network. It is based on the distance vector algorithm and is intended to be subordinate to the existing interior gateway protocol running on the ISPs backbone. It manipulates the route cost calculation so that paths towards the backbone become very cheap and paths away from the backbone become expensive. This forces traffic in the preferred direction unless the backup path ‘shortcut’ is very attractive or the backbone link has disappeared. It is the analysis and development of the fringe routing protocol that forms the content of this ME thesis.</p>

Developing the Fringe Routing Protocol

<p>An ISP style network often has a particular traffic pattern not typically seen in other networks and which is a direct result of the ISP’s purpose, to connect internal clients with a high speed external link. Such a network is likely to consist of a backbone with the clients on one ‘side’ and one or more external links on the other. Most traffic on the network moves between an internal client and the external world via the backbone. But what about traffic between two clients of the ISP? Typical routing protocols will find the ‘best’ path between the two gateway routers at the edge of the client stub networks. As these routers connect the stubs to the ISP core, this route should be entirely within the ISP network. Ideally, from the ISP point of view, this traffic will go up to the backbone and down again but it is possible that it may find another route along a redundant backup path. Don Stokes of Knossos Networks has developed a protocol to sit on the client fringes of this ISP style of network. It is based on the distance vector algorithm and is intended to be subordinate to the existing interior gateway protocol running on the ISPs backbone. It manipulates the route cost calculation so that paths towards the backbone become very cheap and paths away from the backbone become expensive. This forces traffic in the preferred direction unless the backup path ‘shortcut’ is very attractive or the backbone link has disappeared. It is the analysis and development of the fringe routing protocol that forms the content of this ME thesis.</p>

Evaluating Mobility Management Models for Content Forwarding in Named Data Networking Environments

<span>Named Data Networking (NDN) performs its routing and forwarding decisions using name prefixes. This removes some of the issues affecting addresses in our traditional IP architecture such as limitation in address allocation and management, and even NAT translations etcetera. Another positivity of NDN is its ability to use the conventional routing like the link state and distance vector algorithm. In route announcement, NDN node broadcasts its name prefix which consists of the knowledge of the next communicating node. In this paper, we evaluate the performance of mobility management models used in forwarding NDN contents to a next hop. This makes it crucial to select an approach of mobility model that translates the nature of movement of the NDN mobile routers. A detailed analysis of the famous mobility model such as the Random Waypoint mobility and Constant Velocity were computed to determine the mobility rate of the NDN mobile router. Simulation analysis was carried out using ndnSIM 2.1 on Linux Version 16.1. we build and compile with modules and libraries in NS-3.29. The sample of movement of the mobile router is illustrated and our result present the viability of the Constant Velocity model as compared with the Random Way point.</span>