IP Flow Mobility Trigger Mechanism In The Heterogeneous Wireless Networks

With the rapid development of wireless mobile communication technologies, the integration between heterogeneous wireless access networks has become the inevitable trend of future mobile communication. In the integration of heterogeneous wireless networks, how to make full use of the advantages of different access networks, and how to provide seamless continuity for the mobile terminals in a variety of heterogeneous network overlapping area, are hot issues in the research for heterogeneous wireless networks’integration. With the Multi Access Packet Data Network Connectivity and IP flow mobility introduced by3GPP, an operation mode with finer granularity for handover has been provided, at the same time opening a new era for the integration of heterogeneous wireless networks.IP flow mobility makes it possible for a multi-radio UE that is connected to the EPS via different accesses simultaneously, sending and receiving different IP flows through different accesses. Besides, it allows to switching over a single IP flow seamlessly and selectively to a different access, meanwhile keeping the other ongoing IP flows on their original access.Based on the3GPP standard, we introduced the basic concept of IP flow mobility, along with the system architecture, solutions and some signaling procedures. After that, this paper studied the specific implementation mechanism and algorithm in research for IP flow mobility depending on different trigger modes.The existing ANDSF has some drawbacks that it provides inaccurate routing strategies to mobile terminals. An enhanced mechanism with three improvements has been proposed in this paper, providing a better strategy for terminal-based IP flow mobility.Through deeply studying multiple attribute decision algorithms, combined with the characteristics of IP flow mobility, this paper proposes a mechanism for network-based IP flow mobility, which includes a trigger process and a network selection decision algorithm based on each IP flow. Simulation results validate the flexibility and efficiency of our proposed mechanism, which can achieve a better network load balancing as well as a higher user satisfaction.