Impact of Content Popularity on Content Finding in NDN: Default NDN vs. Vicinity-based Enhanced NDN

Named Data Networking allows a consumer to locate a desired content object by its name prefix. By using the best route strategy of the default NDN architecture, an Interest packet is forwarded along a default path indicated by the packet's name to find the content object. However, in this design, the content object can be requested several times and it might be replicated in several nodes that are close to the consumer. Hence, the best path to fetch the content object might not be the default path. We introduce a vicinity-based replica finding scheme to locate the replicas of the content object that are replicated nearby. We have found that the scheme can decrease the delivery time effectively with acceptable overhead costs. However, there are other important factors that can affect the content finding results. For example, locating highly popular content might be faster than finding low popular content depending on the different number of replicas. Cache replacement rates in different populated networks can also affect content finding results. In a high populated network, if the cache size is quite small, the replicas can be replaced easily. It is more difficult to find the replicas that have been replaced. In this paper, we evaluate the proposed scheme against different content popularity levels (from low to high). The results demonstrate that our scheme can perform better compared to NDN default in almost every case.