A Comparison of IP mobility protocols

This paper presentsa detailed comparison, in a comprehensive framework, of Mobile IP and four of the main IP Micr o-mobility protocols.Wefirst describethe global mobility landscapeand point out the important problemsthat must be addressed.Theseare mainly Handoff management,Passi veConnectivity and Pagingsupport, Scalability, Robustness, Quality of Service and Security. Basedon this framework, weexaminein a first stepMobile IP asa Macro-mobility protocol. In a secondstep,wecompare four well-known IP Micr o-mobility protocols: Cellular IP, HAWAII, TeleMIP and EMA. Keywords— Mobile IP, Wirelessnetworks, Cellular IP, Hawaii, EMA, TeleMIP INTRODUCTION BROADBAND WIRELESSNETWORKSarequicklyevolving towardsall-IP networks.Intensi veresearchis currently carriedout to enhanceIP to allow thesenetworks to re-usethe well-known IP mechanisms.In this process,many proposals have beenmadeto enrichIP with the functionalitiesnecessary to managethemobility of users. The mostwidely known of theseproposalsis certainlyMobile IP [1] which is also the oldestone. Mobile IP offers a mechanismallowing usersto changetheir point of attachment in an IP network. Unfortunately, this protocolsuffersfor many weaknesses andthat is the reasonwhy the mobility problemis oftendividedin two parts:macro-mobilityandmicro-mobility. Thedistinctionbetweenthetwo dependsonthescaleof stations movements. The Mobile IP propertiesallow it to be usedas macro-mobilitymanagement protocol. Micro-mobility coversthe management of usersmovements at a local level, inside a given wirelessnetwork. Many solutionshavebeenproposedto managethis typeof mobility within IP networks,they areoftencalledIP Micro-mobility protocols. Sometimesdesignedfor veryspecificissues,theirheterogenous characteristicsandpropertiesdo not allow to easily obtain an accuratepictureof theIP mobility management problems. This paperpresentsa detailedcomparison,in a comprehensive framework, of Mobile IP andfour of the main IP Micromobility protocols. We first describethe global mobility landscapeand point out the importantproblemsthat must be addressed.Basedon this framework, we examinein a first step Mobile IP asa Macro-mobilityprotocol. In a secondstep,we comparefour well-known IP Micro-mobility protocols:Cellular IP [2], HAWAII [3], [4], TeleMIP [5] andEMA [6], [7]. Finally, wepresentourconclusions. Dueto sizelimitations,wecannotaddressall theissuesin this paper . A moredetailedcomparisonis availablein [8]. I . A GLOBAL I P MOBILITY FRAMEWORK This sectionwill focuson thepresentationof a globalmobility landscapeandthemajorissuesfor IP mobility to beinvestigatedwithin this landscape. PierreReinboldand Olivier Bonaventureare with the Infonet group, University of Namur, Belgium. http://www.infonet.fundp.ac.be. E-mail: preinbold,obonaventure@info.fundp.ac.be A. Themobility landscape Thispaperdealswith all-IP networks.Thesearetheexpected futuremobilewirelessnetworks,relyingentirelyon IP: from the mobilestationto thegateway towardstheInternet. Wecall adomaina largewirelessaccessnetwork underasingleadministrationauthority. Suchanetwork is composedof two kindsof machines.Wecall basestation(BS) anequipmentable to communicatedirectlywith themobilenodesvia theradiointerface. In the caseof a CDMA basednetwork, the so-called RadioAccessNetwork(RAN) 1 canbeseenasasingleBSsince the mobility insidethe RAN is managedat the radio layerand is transparento upperlayers.We simply call stationany other network machine.A stationperformingspecialtasksin themobility management will benamedMobility Agent(MA ). We alsoassumethat eachMN hasa HomeNetwork(HN), a domainfrom which it hasobtaineda static2 IP address:its HomeAddress(HA ). We call ForeignNetwork(FN) any other domainwheretheMN canconnect. In suchacontext, wecanreasonablyproposeamodelof what will bethefuturemobility landscape, this is illustratedin figure 1.