Location-based assisted handover for the IP Multimedia Subsystem

The IP Multimedia Subsystem (IMS) is one promising candidate for merging currently deployed circuit-switched and packet-switched voice and data networks into a service-based All-IP architecture. One requirement common to all of these Next Generation Network (NGN) architectures is seamless user mobility - a mandatory feature in order to match with the tremendous success of today's circuit-switched GSM voice networks. In addition most NGN architectures mandate the support for heterogeneous access technologies like WLAN, WiMAX, GPRS/EDGE, UMTS/HSDPA which raises the need for seamless vertical handover between these access technologies. Despite ongoing standardisation and technology developments like Generic Access Network (GAN), Voice Call Continuity (VCC) or Media Independent Handover (MIH, IEEE 802.21) we argue that there is currently no solution to enable seamless and application-transparent vertical handover in the IMS. In this paper, we review vertical handover standards for IMS and propose a novel vertical handover architecture which relies on geographical location and global network information as primary trigger or support for handover initiation. The underlying Session Initiation Protocol (SIP) notification framework - a Terminal-Centric IMS Location Enabler (LE) - is based on the key concept that location information is one specific kind of presence information which is optimally provisioned by mobile devices: the terminal knows best where it is located. The Location Enabler solicits notifications from the mobile device on specific events - e.g., entering or leaving specific geographical areas - thus dramatically reducing the amount of data transferred over the radio interface when compared to tracking or polling. The SIP registration and Re-Invite message flow is used to convey all handover-relevant information between the IMS core network and the mobile terminal. Using accurate access network measurement results the paper presents a performance estimation of the proposed vertical handover architecture with respect to handover duration and required handover area width at different velocities.