NDN Construction for Big Science: Lessons Learned from Establishing a Testbed
Abstract
NDN is one instance of ICN, which is a cleanslate approach that promises to reduce inefficiencies in the current Internet. NDN provides intelligent data retrieval using the principles of name-based symmetrical forwarding of Interest/ Data packets and in-network caching. The continually increasing demand for the rapid dissemination of large-scale scientific data is driving the use of NDN in big science experiments. In this article, we establish the first intercontinental NDN testbed to offer complete insight into NDN construction for big science. In the testbed, an NDN-based application that targets climate science as an example big-science application is designed and implemented with differentiated features compared to previous works on NDNbased application design for big science. We first attempt to systematically address detailed analysis of why or how NDN benefits fit in big science and issues that must be resolved to improve each advantage, mostly based on lessons learned from establishing the NDN testbed for climate science. We extensively justify the needs of using NDN for large-scale scientific data in the intercontinental network, through experimental performance comparisons between classical deliveries and NDNbased climate data delivery, and detailed analysis of why or how NDN benefits fit in big science.
Publication Title
IEEE Network
Recommended Citation
Lim, H., Ni, A., Kim, D., Ko, Y., Shannigrahi, S., & Papadopoulos, C. (2018). NDN Construction for Big Science: Lessons Learned from Establishing a Testbed. IEEE Network, 32 (6), 124-136. https://doi.org/10.1109/MNET.2018.1800088
