Partage d'infrastructures et convergence fixe/mobile dans les reseaux 3GPP de prochaine generation

In a first paper entitled "A Potential Evolution of the Policy and Charging Control/QoS Architecture for the 3GPP IETF-based Evolved Packet Core", FMC and Core Network (CN) sharing aspects are treated simultaneously because it is important that the logical PCC architecture reflects the realities of the industry trends described above.;Following the description of the trends in the communications industry were presented a list of four requirements that enable for a PCC architecture: service convergence (capacity to use a service from any type of access), CN sharing that allows several Mobile Virtual Network Operators (MVNOs) to coexist, the creation of local access network policies as well as efficient micro-mobility in roaming scenarios.;As a second step, two NGN architectures were evaluated upon the requirements mentioned above. This evaluation concluded that a hybrid solution (based on the key features of each architecture but without their respective drawbacks) would offer a very promising foundation for a complete solution.;The proposed solution achieved its goal with a clearer separation of the business roles (e.g., access and network providers) and the introduction of a Network Policy Function (NPF) for the management of the CN. Indeed, the business roles that were defined allow the creation of distinct policy/QoS and administrative domains. The roles become mandatory in infrastructure sharing scenarios. Otherwise, they maintain the compatibility with the actual vertically-integrated operator model; the latter then plays all of the business roles.;Introducing the NPF into the CN enables the CN policy management to be separated from policy management related to subscribers, services and access networks. Additionally, the NPF allows the CN to be shared by multiple Network Service Providers (NSPs) and respect the Service Level Agreements (SLAs) that link the IP Aggregation Network (IPAN) to the NSPs, as well as those that tie the IPAN to the Access Network Providers (ANPs).;A second paper entitled "A Network Policy Function Node for a Potential Evolution of the 3GPP Evolved Packet Core" constitutes an extension of the first paper that extensively described the industry trends, two existing PCC architectures and their characteristics, and finally offered an overview of the proposed solution. On the other hand, the second paper thoroughly describes all of the impacts that the proposal has on the existing 3GPP PCC architecture. Indeed, a significant contribution of this second paper is that it provides an extensive list of potential simplifications that the proposed solution allows.;The main contribution of the second paper is that from now on the proposed solution can be deployed over an existing PCC architecture with a minimum of impacts. Indeed, a small modification to the NPF's reference points enables this enhancement. As a consequence, this enhancement provided a solution that is compatible with both PCC architecture variants, based on either GPRS Tunneling Protocol (GTP) or Proxy Mobile IPv6 (PMIPv6).;A last contribution of the second paper is to demonstrate the NPF's internals when the former is controlling a an IPAN based on tunneling mechanisms such as Multi-Protocol Label Switching (MPLS) or Provider Backbone Bridge-Traffic Engineering (PBB-TE). A traffic engineering process allows traffic flow aggregates to pass around a congested node, to better balance the load between the network elements and make sure that the QoS requirements are respected at all times.;The third paper entitled "A MultiAccess Resource ReSerVation Protocol (MARSVP) for the 3GPP Evolved Packet System" deals with QoS provisioning in FMC scenarios, especially for applications that are not directly supported by the network. As an example, all peer-to-peer applications (such as online gaming) that represent a small fraction of the total peer-to-peer traffic or those that are new and relatively unknown. (Abstract shortened by UMI.).