ANALYTICAL MODELING OF OPTIMAL CHUNK SIZE FOR EFFICIENT TRANSMISSION IN INFORMATION-CENTRIC NETWORKING

As an innovative network architecture, information-centric networking (ICN) is potential to promote network with name-based content retrieval. In ICN, contents are divided into chunks by default. Because chunk is the basic data unit, the size of chunk dramatically affects the transmission efficiency. In this paper, an analytical model is developed to obtain the optimal value of chunk size for content transfer in ICN. Intuitively the larger the chunk size, the better the performance of data transfer. This is because the protocol overhead of each chunk is relatively fixed. However, larger chunk usually leads to larger chunk loss rate and larger response latency. Therefore, the content transfer process is modeled by considering both the protocol overhead and network status. Three typical transmission modes are analyzed, and the expressions for the goodput are derived. By using numerical analysis, the effects of chunk size on goodput under different round-trip times (RTTs), packet loss rates and bandwidths are studied in depth. Based on the model, the recommended values of the chunk size are provided for two deployment approaches in ICN. Finally, experiments are conducted to verify the accuracy of the proposed model.