REPLACEMENT POLICY ALGORITHM PERFORMANCE ANALYSIS OF LEAST RECENTLY FREQUENTLY USED IN NAMED DATA NETWORK

• Main Author: Adi Paramartha Putra, Made
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/39395
• Institution: Institut Teknologi Bandung

Internet user growth increase greatly every year, which affected the network availability and efficiency in order to deliver the content to consumer. Some online entertainment media such as, Youtube, Netflix and Amazon is most accessed website in several last years. Named Data Network (NDN) is the evolution of Content Centric Network (CCN) provide ability to request content to nearest router instead directly request the content to main producer. By using this method, network efficiency improved greatly and greatly reduces delay. NDN router have a capability to cache spesific content into content store based on replacement policy that implement in the content store. In this research, we evaluate Least Recently Frequently Used (LRFU) algorithm, then compared with Least Recently Used (LRU) algorithm and Priority- FIFO algorithm as a replacement policy in NDN router. The simulation will be done by using variation of grid topology and interest rate. Based on the simulation result, LRFU algorithm achieve 5,0% lower miss rate ratio and higher hit ratio compared with LRU and 7,62% with Priority-FIFO. LRFU performance is greater when used in relatively small content store size. If implemented in relatively large size of content store, LRFU performance tends to be the same as LRU and Priority-FIFO replacement policy. In Palapa Ring topology, LRFU algorithm gainned 1,35% lower miss rate and higher hit rate than LRU and 4,42% compared with Priority-FIFO algorithm.