Analysis of NDN repository architectures

Named data networking (NDN) is a newly proposed networking architecture, which a few network scientists believe that it is an appropriate protocol to distribute Big Data files in the network of supercomputing domain. This new scheme is organized to reduce high network congestion resulted from it. In addition to many current NDN applications, such as climate modeling data service, another supercomputing domain, its repository applications also have been developed. However, compared to the ordinary file systems, they now significantly underperform in writing operations. It will be highly likely to prevent scientists and engineers from applying NDN over their networks because it could turn to be as a system bottleneck in the performance critical area. In this paper, writing operations of those repositories will be examined and how performance can differ according to the software architectures placed in them. Finally, we will discuss optimization schemes to make it perform better in a large scaled computing system.