TRUST AND COOPERATION INCENTIVES FOR WIRELESS USER-CENTRIC ENVIRONMENTS Carlos Ballester Lafuente a , Jean-Marc Seigneur a , Waldir Moreira b , Paulo Mendes b , Linas Maknavicius c , Alessandro Bogliolo d , Paolo Di Francesco e a University of Geneva, b Universidade Lusófona Portugal, c Alcatel-Lucent Bell Labs France, d Università degli Studi di Urbino, e Level7 ABSTRACT The flexibility inherent to wireless technologies is giving rise to new types of access networks and it is allowing the Internet to expand in a user-centric way. In this kind of scenario, trust, reputation as well as adequate incentives for cooperation are of vital importance. This paper provides an overview on the state of the art in trust and reputation in wireless environments, identity disambiguation and cooperation incentives. It underlines the current shortcomings in these fields and establishes the relation between these previous concepts in order to form a more robust and reliable Internet, where the end-user is directly involved in the connectivity value chain. KEYWORDS Trust, cooperation incentives, user-centricity, wireless, identity disambiguation. 1. INTRODUCTION Wireless networks today are partially being formed by nodes (e.g. Internet access points, smartphones, femtocells) that are owned and carried by humans. As such, these user-centric networks (UCNs) are giving rise to new Internet architectures, where broadband access is complemented by e.g. Wireless Fidelity (Wi-Fi) clouds, having an active involvement of Internet end-users. This represents a paradigm shift in the Internet evolution, as the user may be in control of parts of the network, in a way that is acknowledged (or not) by Internet stakeholders. In such scenarios where several strangers are expected to interact for the sake of robust data transmission, trust and reputation are of vital importance as these establish a way for the nodes involved in the system to communicate with each other in a safe manner, to share services and information, and above all, to form communities that assist in sustaining robust connectivity models. UCNs have a few characteristics which make them complex to control: i) they are supported both by static, fully dedicated nodes as well as by nodes provided by end-users on-the-fly; ii) in their majority, they are complementary to broadband access but are located on the last hop to the Internet end-user, which is not always accessible from the operator perspective; iii) as some nodes are carried by Internet end-users, their networking composition and organization follows a social behavior, highly affected by human movement features. UCNs are therefore in the category of self-organizing, viral systems, where often anonymous nodes can try to become part of the system and as such, UCNs require that trust and reputation systems/metrics should be attack resistant in order to ensure that the architecture is as robust as possible. Due to their characteristics, a successful deployment of UCN relies on the combination of: i) trust and reputation systems/metrics aiming to create a more flexible system than the usual communication protection schemes (e.g. VPNs); ii) Identify disambiguation aiming to create UCNs resistant to attacks such as Sybil [1] and similar ones [2]. In this paper we consider Sybil attacks to be the ones that are most worrying in UCNs