QoS in Forwarding Strategies for ICN: New Algorithm and Experimental Evaluation

Abstract

Several forwarding strategies (FS) were already proposed to leverage packet delivery in information-centric networking (ICN). However, less attention has been given to FS potential to improve quality of service (QoS) in ICN. This work addresses the impact of FS on the performance of  voice application, video streaming and file transfer protocol (FTP). Particularly, we investigate the performance gain when distinct FS are tailored to these traffic categories using named data networking (NDN), a prominent ICN architecture. This work also proposes a new forwarding strategy, called inverse pending interests (IPI), which can be used for low-priority traffic. Given that distinct FS are simultaneously applied, we also evaluate their interoperability using three scenarios and observing QoS metrics for the applications mentioned above with different priorities for each traffic.
The work was conducted within a multiobjective optimization framework, which allows defining dominant sets of solutions and enriches the analysis when compared to aiming at a single optimal solution. The simulations were performed using the ndnSIM simulator and the results indicate, for example, that multicast and best-route are the most suitable FS for voice application, while the request forwarding algorithm could better serve video streaming. The proposed IPI algorithm is beneficial when low-priority (FTP) traffic is present, and its adoption does not significantly impair other FS.