Design of a user-level naming solution for the future Internet

Overview & Synopsis

Naming is a fundamental element to evolve the current Internet into the next stage. The new host-level and user-level scenarios of the future networks introduce great pressure towards the initial two-level naming systeme of the Internet, which is requested to evolve in order to answer these new requirements. More specifically, special attention should be paid to person-to-person communications and multi-device support in future naming schemes' design. This thesis concentrates on the study of the naming research trends considering to improve future Internet's support to both host-level and user-level requirements. the Identifier/Locator split concept has been widely approved as a crucial solution for current Internet's naming problems. This is why, we first concentrate our workon styduing several Identifier/Locator split solutions. We provide a qualitive overview and a quantitative cost analysis of the proposed approaches. Based on the results that they we have obtained, we emphasize the host-centirc characterof these solutions and we show that they have shortages regarding additional user-level requirements. In order to complement previous research work in the naming area, we present in this thesis a new naming proposal that we call Service-Aware Naming Architecture (SANA). Our solution is built around three key, points. First, we challenge the traditional Identifier/Locator split paradigm bu getting rid of treminals' identifiers and promoting users and services identification. Second, we provide a transparents multi-device support to the network. Ang finally, we make user's session switching between different terminals agnostic to applications and networks. SANA cost analysis results show that our solution has comparative performances with traditionnal Identifier/Locator split solutions. It also outperforms other user-level naming solutions in terms of signaling cost and size of resolution systèms. By covering these aspects intrinsically, we believe that SANA can be considered a significant candidante for future naming systems with respect to user-level requirements.