The effect of scaling the size of Topology and Content Stored on the Named Data Networking

Abstract

NDN (Named Data Networking) as one approach of the Content-Centric Networking (CCN) architecture, proposed by Jacobson, 2006, conceptually, has some advantages compares to the current Internet technology which build based on host-to-host TCP/IP protocol stack. However, technically, we have to examine the NDN approach, beside to prove the concept, also to get as much as able the opportunities of improvement the concept before implementation stage. Some challenges in Large topology such as the effectifity of Content Store size, should be considered first. Some researchers work by top down approach, it means they work by developing sophisticated NDN application to be run on trivial NDN for getting much the requirements for NDN to support the large range of applications. However we work based on bottom up approach, we do the deep examination on the NDN to get much knowledge of the behaviour of NDN itself in the perspective of network discipline by using one particular model of the application. We conduct some simmulation scenarios run on ndnSIM to examine our mathematical model of delay distribution in NDN instead of running the scenarios on a test-bed network that overlay the existing TCP/IP network. An TCP/IP overlay test-bed network will contain some `poisons' of host-to-host TCP communication. The scenarios consist of scalling the size of topology and also the size of content store. The topology choosen is grid topology in a number of size.