In a user-centric approach, the end-user wants the best Quality of Experience (QoE) and the highest context-awareness. Throughout end-user’s spatial mobility, service providers must continuously adapt the end-user’s service session according to his or her new ambient context, without neglecting functional preferences and demanded Quality of Service (QoS). Many spatial mobility management solutions have been conceived to help operators managing the network part of their end-users’ sessions. However, this problem is not yet investigated on the service layer. For this purpose, the authors propose in this paper the concept of semantic handover. This concept acts on the service layer by analogy to the network’s handover mechanisms. It is based on three complementary roles: the initiator, that detects a spatial mobility and initiates the semantic handover; the decider, that decides which services should be replaced in the end-user’s service session and what are their ubiquitous ambient counterparts; and the executor, that executes the decision and adapts the end-user’s service session to the new ambient context. This solution is based on highly scalable distributed coverage zones. Service providers must deploy their ubiquitous services, having the same functionality and an equivalent QoS, in different geographic areas, and they must associate each service to a specific coverage zone. Therefore, when the end-user moves among coverage zones, a semantic handover is launched to guarantee seamless service continuity throughout end-user’s ambient contexts.

Cloud computing has become one of today's hot topics. The major contribution of this Internet-based service delivery paradigm consists in offering computing, storage and network resources able to guarantee information technology externalization. In parallel to this novel trend, cloud users requirements are quickly emerging due to both network and service convergence. Therefore, beyond its externalization solution, cloud must also respond to users needs within this "Next Generation Networks/Next Generation Services" (NGN/NGS) context. Hence, it should offer service personalization for cloud users, take into consideration their mobile context, and guarantee an end-to-end QoS. In this paper, we propose a QoS-based cloud management architecture that overcomes the aforementioned challenges through several mechanisms. First, we surpass mobility and E2E QoS challenges by gathering ubiquitous elements into ubiquity-based virtual communities. Second, we ensure service personalization by proposing a seamless and dynamic service composition based on stateless services. Finally we take into consideration user's ambient context by using location-based virtual communities. Computing models for QoS-aware and location-aware clouds are also provided.

No abstract

No abstract