NDNizing existing applications: Research issues and experiences

A major challenge to potential ICN/NDN deployment is the requirement of application support, namely, applications need to be rewritten or modified in order to run on NDN networks and receive the full benefits. Using a proxy to translate between an application-level protocol and NDN offers a viable solution that balances between development cost and architectural benefits. In this paper, we study on the questions of how to facilitate and incentivize the development and deployment of such protocol translation proxies. We propose to enable existing applications to communicate “off the grid”, i.e., using only local network connectivity without the global Internet, by translating between conventional client-server protocols and NDN. This provides deployment incentives by enabling a useful feature with no or minimal changes to existing applications. By giving the experience of a few protocols, we hope to abstract out some common design patterns that can be reused in developing other application-level proxies. This paper reports our work on IMAP/NDN translation for local email access and XMPP/NDN translation for local group chat. Based on this work, we identify and discuss a number of common design issues including application-level framing, namespace design, application protocol semantics, multiparty synchronization, security and privacy, and real-world deployment challenges.