Cache based on Popularity and Class in Mobile Named Data Network

Abstract

The Named Data Network (NDN) has been proposed as a content-based network that will be replace IP networks because the data communication mechanism is more efficient in large network environments. The problem of traffic load on this IP network has actually been tried to be solved with Content Distribution Network (CDN). However, because CDN is a network that is still based on IP, each request will be addressed to a particular server address. In addition, consumer that always moving causes CDN to not be able to overcome dynamic content changes, while consumer demand will often change direction. NDN shift the host-centric, as in IP network, become content-centric. One important component in NDN is the content storage used for data caching. Because content store has limited capacity, it's important to find a caching strategy to improve the performance of NDN networks. In the previous study, the CAPIC caching technique (Caching based on Popularity and Class ) has been proposed which determines treatment for each service class, but the research is only limited to the fixed network. Meanwhile, user with mobility characteristic will become the majority in the network. The character of the mobile network is very different from the fixed network. Routers that have previously stored the data in the content store cannot always be used, because these routers sometimes can be out of network coverage. In this research, the previous CAPIC that only work in the fixed environment is modified to be able to used in the mobile network. This research is conduct to analyze the CAPIC performance in the mobile environment for various conditions. The test results showed CAPIC provided a greater network cache hit ratio than the sharing caching scheme of around 3%, with the first service class performance reaching a maximum 10% compare to sharing scheme. CAPIC also provides a better path stretch for content service classes with a smaller request rate.