Q3M – QoS Architecture for Multi-user Mobile Multimedia Sessions in 4G systems [chapter]

Eduardo Cerqueira, Luis Veloso, Augusto Neto, Marília Curado, Paulo Mendes, Edmundo Monteiro
2007 Lecture Notes in Computer Science  
Fourth generation systems (4G) will provide multimedia group communication sessions to multiple mobile users with distinct requirements. This way, it is expected the control of the quality level, connectivity and ubiquitous access for multi-user multimedia sessions across heterogeneous and mobile networks with seamless capability. This paper analyses the requirements of a control architecture to provide Quality of Service (QoS), connectivity and seamless handover management for multi-user
more » ... ns in 4G systems, and introduces the QoS Architecture for proposal. In addition, simulation results present the efficiency of this approach concerning the session setup time and packet losses during handover. Keywords: Multi-user Multimedia Sessions, Seamless Mobility, Quality of Service, Heterogeneous Mobile Networks. reduce operational costs and enhance efficiency in the usage of network resources, where these requirements are satisfied by an all-IP 4G mobile system. In addition to the access technologies heterogeneity, the 4G systems are expected to be heterogeneous also at the network layer, due to the diversity of connectivity control mechanisms and different address realms. For instance, in terms of unicast mobility management, the use of Hierarchical Mobile IP (HMIP) allows users to use a sequence of two global IP addresses, inside and outside each access-network. In the field of IP multicast, which is the most suitable technology to distribute multi-user sessions, there are already several address realms associated with multicast models such as, the Any-Source Multicast (ASM), Source-Specific Multicast (SSM) and Small-Group Multicast (SGM). However, to allow the distribution of multi-user sessions throughout heterogeneous access and transport technologies, it seems to be mandatory to cluster homogeneous network devices into domains (named clusters in this paper) and to create of open interfaces between such clusters and existing standards. Therefore, a control architecture should be built as modular and decentralized as possible. The modularization facilitates the inclusion of emerging technologies, while the decentralization permits a higher scalability. Both of these characteristics might be sustained by an edge-networking approach in which the functionality of each cluster is controlled by a group of organized edge devices. Since each cluster may implement a different connectivity scheme (e.g., IP unicast, IP multicast, Multimedia Broadcast Multicast Service (MBMS) or any type of Layer 2 connectivity), the architecture must control the multi-user session connectivity cluster by cluster. This functionality allows the session continuity independent from the connectivity scheme offered inside or between clusters. Hence, each multi-user session must have a global identifier, independent from hosts location or IP address. Furthermore, in order to support emerging mobile-aware sessions, to increase the satisfaction of the users and to avoid the interruption of sessions, an architecture must provide ubiquitous access, QoS and mobility control across heterogeneous clusters with no perceived service degradation for the user. From the QoS point of view, different QoS models (e.g., such as Differentiated Service (DiffServ), IEEE 802.11e and IEEE 802.16) can be implemented by clusters to provide QoS assurances for multi-user sessions along end-to-end heterogeneous paths. Hence, inside or between clusters, QoS mapping procedures must be done to map the requirements of the session into the appropriated service class. Since QoS mapping alone is not sufficient to assure the quality level of sessions, due to the use of different QoS models, class of services with distinctive configurations and available bandwidth capacities, QoS adaptation schemes must be used to avoid the session blocking and to keep those sessions with acceptable quality level, independently of the movement of users or even re-routing events (caused by a link or network agent failure). From the mobility point of view, an architecture must support inter and intra-cluster handovers with seamless capability. For instance, by using caching and buffering mechanisms to reduce packet losses during handover and by reserving network resources in advance to allow a faster session re-establishment. Summing up, an IP-based control architecture for multi-user multimedia sessions must allow senders to offer their content and receivers to access them ubiquitously. It must also provide open interfaces, QoS, connectivity and seamless mobility control.
doi:10.1007/978-3-540-75869-3_4 fatcat:vguyc4xnnzg4lgiyt4urnqkb7a