SC13 Denver, CO

The International Conference for High Performance Computing, Networking, Storage and Analysis

Evaluating Named Data Networking For Large Scientific Data.

Student: Susmit Shannigrahi (Colorado State University and Lawrence Berkeley National Laboratory)
Supervisor: Alex Sim (Lawrence Berkeley National Laboratory)

Abstract: With growing volume of scientific data, data management and retrieval are becoming increasingly complicated. Named Data Networking (NDN) is a potential next generation Internet architecture. In this preliminary work, we show that Named Data Networking (NDN) offers unique properties that can reduce complexities involving large scientific data. We discuss these built in properties of NDN for reducing the need for catalogs, complex middleware or applications. We then support these observations with results from experiments done using the ESNet testbed and real scientific data. We have used a comparatively small dataset and simple testbed topology to demonstrate NDN's capabilities. We plan to extend this work for real scientific analysis involving larger datasets and different network topologies. This poster focuses on demonstrating how NDN is useful for large scientific data. It also shows NDN's ability to reduce middleware and complex applications and at the same time, offer improved performance.

Poster: pdf
Two-page extended abstract: pdf

Poster Index