Proposition d'une approche opportuniste de gestion de la Mobilité dans les réseaux communautaires sans fil Basée sur le modèle de Markov

Abstract

Nowadays, network and mobile radio technology has made considerable progress. We are witnessing the deployment of different telecommunication network standards such as GSM, UMTS WiMAX and wireless local networks such as WiFi (IEEE802.11x). Community wireless networks have also rapidly developed with the arrival of WiFi thanks to their simplicity, speed and low cost of spreading out. Ever since, they are an interesting alternative to traditional local networks. Mobility management in wireless networks is now a real challenge in this type of network with the proliferation of mobile devices. A mobile user can change its network access point by moving, without interrupting the current service session, that is, to undergo handoffs. Our study is based on the high layers of the OSI model (network, application and transport). The Network layer IP Mobile Protocol is by IETF-standardized protocol that allows the user to maintain current communications and stay connected to the Internet while seamlessly masking the network change. The design of Mobile IPv6 was based on the experiences gained from the development of Mobile IPv4 and the new opportunities offered by IPv6, such as the larger number of addresses, self-configuration and security mechanisms. Although IPv6 Mobile protocol solve addresses the problem of mobility in IPv6, this protocol can not support applications in real time that are sensitive to delay or loss of packets. Indeed, in the standard version of Mobile IP, the new location of a mobile is always reported to its home agent. The latter is thus notified of all the movements of the mobiles he manages. The loss of some Packets during handovers can be important since the registration process is long, especially if the Home Agent-HA is at the other end of the world. The duration of a handover can reach several seconds in the current Internet. Mobile IP improvement, such as (HMIP, hierarchical MIP), (FMIP, Fast Handover for MIP) is the reduction of handover latency and the loss of packets during a Hanover. The basic idea of the Fast Handover in Mobile IP [10] is to allow the mobile node to obtain its new temporary address before performing the Hanover to the new subnet so that it can immediately communicate when the connection with its new access router is established. This requires an anticipation of movements. At the application layer level, SIP (Session Initialization Protocol) is control protocol application layer that can establish, modify, and terminate multimedia sessions, such as telephone communications over the Internet. The SIP protocol provides location management for terminal mobility. The basic SIP protocol does not provide transparent management of handoff [11]. Thus, the SIP session ends when the mobile changes IP network because the underlying TCP / UDP connection addresses will no longer be valid for the new IP addresses. However, SIP protocol can be used in conjunction with other Handover Management schemes: Mobile IP (MIP); Cellular IP (CIP) mobile stream control transmission protocol (mSCTP). Among the mobility protocols at the transport level, our concern is in the SCTP / mSCTP protocols given the devices they offer. SCTP implements two new mechanisms, Multihoming and multisteaming that make it very reliable for mobility management and insignificant packet losses. The Multihoming mechanism allows the management of IP addresses, the control of borrowed addresses (or paths) and the transfer of data in an association. The modeling of the mSCTP mobility protocol using the Markov chain allowed us to choose the parameterδ (fixed or variable threshold) allowing the free movement of the mobile in the cells of high probability cells retained by the protocol. In addition, in a context where mobile users are engaging more and more sensitive applications in time, mobility management remains a priority issue, requiring quality of service (QoS). Aggregation functions have allowed us to make a quantifiable judgment on several intercellular transitions that can increase the quality of service for audio and video applications. To reach a consensus on these judgments, two aggregation functions have been proposed: the arithmetic and quasi-linear average. The choice is focused on the quasi-linear average thanks to the parameter wi (power of the signal), which represents the weight of the handover and also depends on the altitude between two antennas in interference, and also the geographical obstacles. Keywords: