CC-NIE Integration: Supporting Climate Modeling Over Named Data Networking (NDN)

Colorado State University, Fort Collins CO

Investigators

Abstract

Named Data Networking (NDN) is an instance of Information Centric Networking (ICN). Unlike IP, NDN focuses on content (the what) rather than hosts (the where). In NDN users simply request content by name rather than its location, and the network locates and retrieves signed content from anywhere, including the publisher, a network cache or a nearby user. The ability to specify content by name and retrieve the nearest copy is of great benefit to Big Data applications by seamlessly locating related data and speeding up retrieval of large datasets, especially in collaborative environments. In this project, networking researchers at the Computer Science and Atmospheric Sciences departments at Colorado State University will work together to integrate NDN with the Global Cloud Resolving Model (GCRM) application. GCRM simulates the global atmosphere using a grid fine enough to crudely resolve the larger individual clouds. This application typically generates terabyte to petabyte scale datasets. GCRMs are currently under development at several modeling centers in the U.S. and abroad. Project contributions include a seamless integration of NDN with researchers' work flow through integration with the file system, versioning control, the design of appropriate name spaces to facilitate seamless data discovery, building efficient collaboration and sharing environments and ensuring data integrity. To ensure smooth transition to the new environment, the project builds translators between existing and NDN-appropriate formats. Instrumented NDN routers are deployed across three campuses to support testing and engagement in a continuous feedback loop with researchers to improve the network and the tools.