Evaluating CCN multi-path interest forwarding strategies
Résumé
This work addresses the performance evaluation of Content Centric Networks (CCN). Focusing on a realistic YouTube-like catalog, we conduct a thorough simulation study of the main system performance, with a special focus on multi-path interest forwarding strategies but thoroughly analyzing the impact of several other ingredients -- such as network topology, content popularity, caching decisions and replacement policies.
Summarizing our main results, (i) catalog and popularity settings play by far the most crucial role (ii) the impact of the strategy layer comes next, with naive forwarding strategies playing against CCN efficiency, (iii) simple randomized caching policies perform almost as well as more complex ones, (iv) the impact of the topology is limited. Hopefully, our thorough assessment of scenario parameters can assist and promote the cross-comparison in the research community -- for which we also provide our CCN simulator as open source software.