Mobility management with distributed mobility routing functions

• Published in: Telecommunication systems 2015-06, Vol.59 (2), p.229-246
• Main Authors: Ernest, Petro Pesha, Chan, H. Anthony, Xie, Jiang, Falowo, Olabisi Emmanuel
• Format: Article
• Language: English
• Subjects: Anchors; Artificial Intelligence; Business and Management; Computer Communication Networks; Computer networks; Internet; IT in Business; Management; Mathematical models; Networks; Packets (communication); Probability Theory and Stochastic Processes; Service introduction; Simulation; Telecommunications systems; Traffic engineering; Traffic flow
• ISSN: 1018-4864; 1572-9451
• DOI: 10.1007/s11235-014-9958-4

Currently, IP mobility management solutions standardized by IETF rely on a centralized and static mobility approach, whereby a central mobility anchor manages mobility signaling and data traffic forwarding. All data traffic sent to a mobile node (MN) traverse the central mobility anchor regardless the current point of attachment of the MN in the Internet. Consequently, the packets follow a non-optimal path via the central mobility anchor which is typically located in the MN’s home network, and result to increase in packet delivery latency. This paper discusses a distributed mobility management mechanism that releases the burden of the central mobility anchor by splitting the mobility management functions of the mobility anchor and co-locating the mobility routing function at the gateway of each sub-network. This mechanism optimizes the data path of the MN that has moved to visited networks, which reduces the packet delivery latency. Moreover, the mechanism is extended to achieve seamless handover when the MN moves between visited networks. Through simulation under NS-2, the performance of the packet delivery latency, handover delay and packet loss is studied and analyzed with respect to varying network load and increasing distance of the MN’s home network from the visited networks. Simulation results show the improvement in packet delivery latency as compared to using non-optimized paths. Moreover, the results also verify the achievement of seamless handover when an MN moves between visited networks.