CACHING POLICY FOR NDN NETWORK

• Main Author: Situmorang, Hamonangan
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/75247
• Institution: Institut Teknologi Bandung

Content Centric Network (CCN) architecture has changed communication paradigm from host oriented to content oriented. One of implementation approach of CCN architecture is Named Data Networking (NDN). However forwarding function of NDN effectively is depended by capability and capacity of caching system. CEE (Caching Everything Everywhere) in NDN requires unlimited cache capacity in router. Therefore, practically NDN is not realistic approach. It requires some efforts to improve forwarding and caching strategy in NDN. The expected proposed strategy should solve the practical challenge and also provide maximum throughput and minimum delay environment. Performance improvements which offered by NDN are also not able to provide the equality service for two or more traffic with significantly different intensity rate. The inequality service means disparity of hit ratio which tend to influence delay performance. Traffic with high intensity rate exploits the capacity of content store. This phenomenon clearly showed while simulation is taken to compare the performance between NDN and TCP/IP on Inherent (Indonesia Higher Education Network). Even though the average of NDN decreases, the unfair of service between regions is showed. Based on simulation, cache replacement strategy is the important aspect to be modified for providing the universality of service in NDN. In this research, numerical analysis has been conducted to estimate performance improvement opportunity based on hit ratio measurements. Finally, our research has simulated caching policy based on LRFU and equality aware LRFU algorithm. Based on the result it is concluded that LRFU algorithm performance is influenced by model of interest generator. Disparity of frequency between traffic flows make domination from one traffic flow to the other can be reduced by equality aware LRFU algorithm.