Interest Classification on Named Data Network Using the Supervised Learning Method

Abstract

A Named Data Network (NDN) is the future network that changes the previous data communications paradigm. Content-based data is delivered over the NDN network. Consumer requests content via the NDN network in the form of names. Next, the content name is sent to the network and forwarded through network nodes to the destination. Several strategies in the NDN network have an important role, one of which is the forwarding strategy. A forwarding strategy is a set of rules that are used to determine the next hop of interest. These rules are designed to ensure that interest is forwarded in the most efficient manner possible. However, implementing the forwarding strategy without knowing the appropriate face for forwarding the interest to the producer or nearest node will result in high delay and packet drops, affecting QoS parameters and NDN network performance. To overcome this issue, this study applies supervised learning to classify the content requests (interests) requested by consumers. This classification technique is necessary to classify consumer-requested interests supported by several related variables. The outcomes of the conducted research show that raw data from mini-NDN output can be processed and transformed into a set of data. This data is used to train a classification model using supervised learning. In the scenario of 9 NDN nodes with varying numbers of interests, distributed uniformly and according to Zipf's law, The Random Forest model is a good choice for this task with an accuracy rate of 86.2% and an error rate of 14.8%.