Mobility Modeling and Handoff Analysis for IP/MPLS-Based Cellular Networks Rami Langar ¶‡ , Nizar Bouabdallah τ , Samir Tohme ‡ , and Raouf Boutaba ¶ ( ¶ ) School of Computer Science, University of Waterloo ; 200 University Ave. W., Waterloo, ON, Canada ( ‡ ) GET-Telecom Paris, LTCI-UMR 5141 CNRS ; 46, rue Barrault, 75634 Paris, France ( τ ) INRIA, Campus universitaire de Beaulieu ; 35042 Rennes Cedex, France E-Mail: rami.langar@enst.fr ; nizar.bouabdallah@inria.fr ; rboutaba@bbcr.uwaterloo.ca Abstract— One of the major challenges for the wireless net- works is related to efficient mobility management issue. In this paper, we propose a new micro-mobility management scheme, called Micro Mobile MPLS, that supports both mobility and quality-of-service (QoS) management in cellular networks. Our proposal includes two protocol variants. In the first variant, called FC-Micro Mobile MPLS, the forwarding chain (FC) concept is provided to track efficiently the host mobility within a domain. This concept fits mobile nodes (MNs) with high mobility rate. The second protocol variant, called Master Forwarding Chain (MFC)-Micro Mobile MPLS, aims to reduce the total signaling cost by controlling the number of registration updates with the root of the domain. In order to assess the efficiency of our proposals, the aforesaid protocols are compared with respect to the existing solutions. To achieve this, we develop analytical models to evaluate both registration updates and link usage costs. Numerical and simulation results show that the proposed mechanisms can significantly reduce the registration updates cost and provide low handoff latency and packet loss rate under various scenarios. I. I NTRODUCTION Future wireless networks are expected to provide IP-based coverage and efficient mobility support with end-to-end Qual- ity of Service (QoS) requirements. Mobile IP [1], which is a standard proposed by the Internet Engineering Task Force (IETF), can serve as the basic mobility management in IP- based wireless networks. However, it presents several draw- backs such as the long handoff latency and the large signaling load for frequent registration updates. In this regard, several enhancements to Mobile IP for MNs with frequent handoffs have been studied in [2]– [7]. Specifically, authors in [2] propose a distributed dynamic regional location management scheme for Mobile IP to reduce the overall signaling cost. They assume that every Foreign Agent (FA) has the functionality of a FA and Gateway Foreign Agent (GFA). However, this assumption is not realistic and there is no provision for end-to-end QoS support. Authors in [3] showed the limitations of this approach due to its limited applicability. In [4], a fast handoff mechanism for Mobile IP, called FMIP, is proposed. This approach has a significant effect on the performance of real-time and QoS sensitive applications. However, the location update cost in FMIP can be excessive, especially for the mobile nodes with relatively high mobility and long distance to their Home Agents (HAs). On the other hand, the notable benefits of MPLS [8] in terms of QoS, traffic engineering and support of advanced IP services, such as virtual private networks, inspire some works to use this technology in the wireless infrastructure [9] – [14]. In view of this, [9] proposes a scheme to integrate the Mobile IP and MPLS protocols. This scheme, called Mobile MPLS, aims to improve the scalability of the Mobile IP data forwarding process by removing the need for IP-in- IP tunneling from the HA to the FA using Label Switched Paths (LSPs). However, such a scheme suffers from the non- applicability to micro-mobility, as the scope of Mobile IP is more shifted towards the global mobility. In [10], an enhanced label edge router (LER) called the label edge mobility agent (LEMA) is introduced to support intra-domain mobility using LSPs redirection. The scheme is scalable and suitable for QoS support. However, the algorithms for choosing the LEMAs for a particular MN are quite complex. H-MPLS [11] and several other schemes ( [12] [13] [14] ) try to ameliorate the performance of Mobile MPLS [9] by using different architectures. A Foreign Domain Agent (FDA) is introduced into each MPLS domain to support intra-domain mobility. However, these works have not taken into account the fact that the signaling delay for the location update could be very long, which may cause service disruption for real-time services and will result in increasing the registration updates cost, the loss of a large amount of in-flight packets and the degradation of QoS. Note that the in-flight packets are the packets possibly lost during the handoff period. In addition, with high mobility rate, the system performance is critically affected by frequent registrations with the FDA, resulting in excessive signaling traffic and long service delay. To overcome these limitations, we propose in this paper a new protocol called Micro Mobile MPLS. Our proposal supports two protocol variants. In the first variant called FC- Micro Mobile MPLS, the forwarding chain (FC) mechanism, which is a set of forwarding path, is provided to track efficiently the host mobility within a domain. The forwarding chain mechanism fits the wireless environment with high mobility rate, where packets must be quickly redirected to their new locations. On the other hand, the second protocol variant, called Master Forwarding Chain (MFC)-Micro Mobile MPLS, aims to reduce the total signaling cost by controlling the number of registration updates to the root of the domain. To gauge the effectiveness of our proposed mechanisms, we derive analytical expressions of both registration updates and link usage costs. Numerical and simulation results show that our proposals can significantly reduce the registration updates cost and also provide low handoff latency and packet loss rate when compared to the existing schemes (FMIP [4], MIP-RR [5], Mobile MPLS [9], H-MPLS [11]) under various scenarios. The remainder of this paper is organized as follows. Section II introduces our proposed architecture along with a detailed description of the above mentioned protocol variants. Section III describes the system model used to evaluate the perfor- mance of the proposed schemes. In section IV, we develop analytical models to derive the signaling cost function of registration updates and the link usage for all underlying pro- tocols. Numerical and simulation results are given in section