Mobility Management in Heterogeneous Wireless Networks - Larim

638

IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, VOL. 24, NO. 3, MARCH 2006

Mobility Management in Heterogeneous Wireless Networks Abdoul Djalil Assouma, Ronald Beaubrun, Member, IEEE, and Samuel Pierre, Senior Member, IEEE

Abstract—In heterogeneous wireless networks, mobile users are able to move from their home networks to different foreign networks while maintaining access capability to their subscribed services, which refers to global mobility. One of the key challenges in global mobility management is intersystem location management, which consists of keeping track of mobile users who roam into foreign networks. This paper presents an overview of mobility management in heterogeneous wireless networks and introduces a scheme which improves location management efficiency in terms of total signaling costs and intersystem paging delay. More specifically, cost reduction reaches about 50% when comparing the proposed architecture with conventional architectures. Index Terms—Global mobility, heterogeneous wireless network, intersystem location management, paging delay, signaling cost.

I. INTRODUCTION

C

URRENT wireless networks allow mobile users carrying multimode terminals to move from their home networks to different foreign networks while being able to fully access their subscribed services [1]–[11], [21], [22]. This refers to global mobility and service portability. In this context, locating a user or allowing him to access his services requires interoperability between several fixed and mobile subsystems that do not necessarily implement the same technology, which may increase the signaling traffic and decrease the network performance. In a mobile network, the service area is divided into location areas (LA), and each LA covers several cells [13]–[15]. In principle, whenever a mobile terminal enters a new LA, it must update its location information with the network, which allows the network to know exactly its current LA at any time [17], [31]. Implementing LA-based methods for mobility management requires the use of a home location register (HLR) and several visitor location registers (VLR) [12], [24]. In fact, when a mobile user first subscribes to wireless services, a permanent record of his profile is created in the HLR. Since this user may move from one LA to another, his current location is usually maintained in a VLR and must be identified before the setup of any connection. Heterogeneous wireless networks consist of several subsystems which use different protocols and access technologies [29]–[33]. Each subsystem divides its service area into a Manuscript received January 16, 2005; revised May 25, 2005. This work was supported in part by NSERC, FQRNT, and the NSERC-Ericsson Chair. This paper was presented in part at the IEEE WiMob 2005. A. D. Assouma and S. Pierre are with the Mobile Computing and Networking Research Laboratory (LARIM), Department of Computer Engineering, École Polytechnique de Montréal, Montréal, QC H3C 3A7, Canada (e-mail: [email protected]; [email protected]). R. Beaubrun is with the Mobile Network Applications Research (MONARC) Group, Department of Computer Science and Software Engineering, Université Laval, Laval, QC G1K 7P4, Canada (e-mail: [email protected]). Digital Object Identifier 10.1109/JSAC.2005.862407

Fig. 1.

Conventional procedure for intersystem registration.

number of location areas. In this context, intrasystem roaming limits users’ movements between LA within a particular network, whereas intersystem or global roaming considers subscribers’ movements between subsystems using different technologies. In the same vein, intersystem location update concerns updating the location information on a mobile terminal performing intersystem roaming; whereas, intersystem paging is aimed at searching for the called terminal roaming between different service areas. It turns out to be important to develop a strategy for global mobility management which facilitates interoperability between the different subsystems. The rest of the paper is organized as follows. Section II presents and analyzes several well-known approaches proposed for global mobility management. Section III introduces a new mobility management scheme, as well as proposed procedures for intersystem registration, updating, and paging processes. Section IV presents performance analysis of the proposed architecture, in terms of total signaling costs and intersystem paging delay; whereas, Section V presents some concluding remarks. II. BACKGROUND AND RELATED WORK Recently, a number of methods have been proposed to analyze the impact of global roaming on network performance [4], [16], [18]–[20], [23], [24], [28]. Among them, conventional methods do not use any equipment to interconnect subsystems using different technologies, as intersystem location management is made by means of signaling messages directly exchanged between the subsystems. More specifically, suppose a mobile user wants to move from subsystem to subsystem . In this case, intersystem registration is possible if there is an agreement between operators of both subsystems. If so, the registration process may be complex since the mobile user must have an entry within the HLR of the visited subsystem (i.e., HLR ). Such a process is illustrated in Fig. 1 and described as follows. 1) The user sends a registration request to MSC/VLR . 2) The MSC transmits this request to HLR .

0733-8716/$20.00 © 2006 IEEE

ASSOUMA et al.: MOBILITY MANAGEMENT IN HETEROGENEOUS WIRELESS NETWORKS

Fig. 2.

Conventional procedure for intersystem location update.

3) HLR sends a query to HLR to obtain its authorization. 4) HLR updates the user profile and sends a confirmation message to HLR . 5) HLR sends a confirmation message to MSC/VLR . 6) MSC/VLR registers the user profile in its database and sends a response to HLR . 7) HLR sends a confirmation message to HLR . At the end of the process, the mobile user may be served by subsystem . Moreover, intersystem location update concerns movements from an old MSC/VLR to a new one within the visited subsystem [25]–[27], [34]. In the context of conventional methods, the procedure for such a process is illustrated in Fig. 2 and described as follows. 1) The mobile terminal sends a service request to the new MSC/VLR . 2) The new MSC/VLR transfers such a request to HLR . 3) HLR sends an update message to HLR in order to keep subsystem informed of the user movement. 4) HLR sends a confirmation message to HLR . 5) HLR transmits a cancellation message to the old MSC/VLR . 6) The old MSC/VLR sends a confirmation message to HLR . 7) HLR transfers the confirmation message to the new MSC/VLR which creates an entry for the user. Furthermore, with conventional methods, intersystem paging involves searching a mobile user in both adjacent subsystems. More specifically, HLR looks for the user in subsystem first. If this user is not found in subsystem , he will be searched for in subsystem . This scenario is illustrated in Fig. 3 and decomposed as follows. 1) A call arrives at HLR , which sends a message to the last MSC/VLR that registered the user in subsystem . 2) That MSC/VLR replies to HLR that the mobile user has not been found in its LA. 3) HLR sends a message to all adjacent HLR to ask for the current user location. 4) HLR replies that the mobile user is in its service area. 5) HLR transmits a request message to MSC/VLR (which serves the mobile user). 6) MSC/VLR sends a temporary location directory number (TLDN) to HLR . 7) HLR transfers the TLDN to HLR which forwards it to the calling MSC/VLR.

639

Fig. 3. Conventional procedure for intersystem paging.

More recently, approaches for global mobility management have used specialized equipment to facilitate interoperability of subsystems using different technologies [4], [5], [18], [19], [32]. In particular, the approach presented in [4] uses a Wireless INterworking Gateway (WING) in order to reduce signaling traffic in the context of global mobility management. However, it has not taken into account the delay generated during the handoff process, which may result in a degradation of the network quality of service. A similar approach presented in [19] is based on the user profile to manage global mobility. This approach implements the intersystem update process by using dynamic regions called boundary location areas (BLA). Each BLA is controlled by a boundary intersystem unit (BIU) and is evaluated in a function of the speed and the quality of service required by the user application. Intersystem paging is based on a boundary location register (BLR) which is connected to each MSC/VLR of both subsystems and contains data from users who cross a BLA. It has been proven in [31] that this approach may significantly reduce the signaling costs for high-mobility users. However, it does not specify how the registration process is set up when the mobile user turns on his terminal only after arriving at the visited subsystem. Furthermore, since a BLR is connected to each MSC/VLR of both subsystems, an intersystem paging process requires the network to consult the HLR of both involved subsystems, which may increase the signaling traffic generated during the process. III. PROPOSED LOCATION MANAGEMENT SCHEME This section introduces an architectural model which improves network performance in the context of intersystem location management as well as the procedures for intersystem location registration, updating, and paging management. A. Basic Idea and Principle In the context of global mobility, it is imperative to have an architecture which not only guarantees the connections for mobile terminals roaming between heterogeneous networks but also supports service portability between wireless subsystems. Such an architecture may be possible through the use of a specialized equipment called Location Register and INternetworking Gateway (LR-ING), which interconnects the HLR of adjacent subsystems. This interconnection is different than that presented

640


Fig. 4. Interconnection of subsystems i and j with an LR-ING.

Fig. 5. Procedure for intersystem registration with LR-ING.

in [19], as it enables us to reduce not only the financial link cost but also the total signaling cost resulting from global mobility. Fig. 4 gives an example of the interconnection of two adjacent subsystems and in accordance with the proposed architecture. In this case, the LR-ING (denoted LR-ING ) facilitates interoperability between subsystems and by converting signaling information exchanged between them. Furthermore, it contains a database which enables us to collect and register user profiles as well as information related to the sessions of users who change subsystems. As a result, during the paging process, the LR-ING uses its database to retrieve the mobile terminal in the visited subsystem, which enables us to decrease the paging delay during the intersystem handoff. Moreover, the proposed approach defines a boundary location area (like in [32]) to guarantee that the update procedure, as well as authorization, and resource reservation are executed before entering the visited subsystem. This area is dynamic, i.e.,

configurable according to the user profiles. To evaluate the approach performance, we define the following parameters: transmission cost from an MSC/VLR to HLR; transmission cost from HLR to LR-ING; access cost to MSC/VLR; access cost to HLR; access cost to LR-ING. B. Procedures for Intersystem Registration, Handoff, and Location Update Consider two adjacent subsystems and . When a subscriber of subsystem turns on his terminal for the first time in subsystem , he has to register in . As shown in Fig. 5, the registration process is triggered by the MSC/VLR of the visited subsystem, i.e., MSC/VLR . In fact, when MSC/VLR detects the presence of an unknown user in its LA, it sends an authentication


Fig. 6.

641

Procedure for intersystem handoff with LR-ING.

request to HLR . Using the global title translation (GTT) procedure, HLR recognizes that the user comes from subsystem and notifies it to LR-ING . The latter updates its database by creating an entry for the user profile then transfers this information to HLR . More explicitly, the registration process is executed as follows. 1) The user sends an authentication request to MSC/VLR . 2) MSC/VLR transfers this request to HLR . 3) HLR transfers the same request to LR-ING . 4) LR-ING registers the subscriber profile in its database then sends an update request to HLR . 5) HLR updates its database and sends a confirmation message to LR-ING . 6) LR-ING sends a response to MSC/VLR so that it registers the subscriber in its database. At the end of the process, the user may be served by subsystem which uses the LR-ING to keep track of each mobile user coming from other subsystems. Moreover, when a subscriber from subsystem moves into subsystem during a communication, an intersystem handoff is engaged to allow him to access his services from subsystem . More specifically, when a mobile terminal enters the boundary location area, it sends an intersystem update request to LR-ING . The latter translates the signaling message into a format comprehensible by subsystem , authenticates the user identity, sends an update request to subsystem , and updates the user profile in its database. Such operations are illustrated in Fig. 6 and decomposed as follows. 1) The mobile terminal notifies MSC/VLR that it is moving into subsystem . 2) MSC/VLR transfers the request to HLR . 3) HLR sends the same request to LR-ING . 4) LR-ING creates an entry to the mobile user in its database then sends a connection request to HLR . 5) HLR sends to LR-ING the parameters required to maintain the connection, such as bandwidth and available channels.

6) LR-ING transfers such information to MSC/VLR . 7) LR-ING transfers such information to the mobile terminal via MSC/VLR . In the same vein, when a subscriber moves throughout the LA of subsystem (i.e., from an old MSC/VLR to a new one), he has to update his location. In the context of the proposed architecture, the procedure for intersystem location update is illustrated in Fig. 7 and consists of the following operations. 1) The mobile terminal sends a service request to the new MSC/VLR . 2) The new MSC/VLR transfers the request to HLR . 3) HLR transfers the same message to LR-ING . 4) The new MSC/VLR sends a cancellation message to the old MSC/VLR in order to update its database. 5) The old MSC/VLR sends a notification message to the new MSC/VLR . 6) LR-ING sends a notification message to the new MSC/VLR . C. Procedure for Intersystem Paging Intersystem paging requires determining the current location area, regardless of the subsystem the mobile user is located at. In this context, two possible scenarios have been identified for the user location in accordance with the proposed architecture. Scenario 1: Called Terminal is Located at BLA of Home Subsystem: This scenario represents the situation where the called terminal makes an intersystem registration request, while still being located at the boundary location area in its home subsystem. In this context, the procedure for intersystem paging is illustrated in Fig. 8 and consists of the following steps. 1) A call arrives at HLR , which realizes that the mobile user has already made an intersystem registration request. Thus, it transfers the request to LR-ING . 2) LR-ING notifies HLR that the user is still located at subsystem . 3) HLR sends a signaling message to MSC/VLR , which controls the current user location.

642


Fig. 7. Procedure for intersystem location update with LR-ING.

naling costs obtained from the proposed architecture with those obtained from existing models. A. Total Signaling Costs and Intersystem Paging Delay

Fig. 8.

Procedure for intersystem paging: Scenario 1.

Fig. 9.

Procedure for intersystem paging: Scenario 2.

4) MSC/VLR finds the user location, assigns a TLDN to the called terminal, and transfers this information to HLR , which forwards it to the calling MSC/VLR. Scenario 2: Called Terminal is Located at Visited Subsystem: In this context, the procedure for intersystem paging is illustrated in Fig. 9 and consists of the following steps. 1) A call arrives at HLR which transfers the message to LR-ING . 2) LR-ING notifies HLR that a user from subsystem has to receive a call. 3) HLR sends a signaling message to MSC/VLR , which controls the user location. 4) MSC/VLR assigns a TLDN to the terminal and transfers this information to HLR via LR-ING . IV. PERFORMANCE EVALUATION AND NUMERICAL RESULTS To evaluate the performance of the proposed architecture, we have estimated the total signaling costs and paging delay for intersystem registration, handoff, and location update, as well as paging procedures. Afterwards, we have compared the total sig-

To evaluate the total signaling costs, we have defined the following parameters: number of calls per unit of time to mobile terminal; number of times a mobile terminal changes location areas per unit of time; number of adjacent subsystems to a given subsystem; cost for paging a terminal using LR-ING and considering scenario 1; cost for paging a terminal using LR-ING and considering scenario 2; cost for paging a terminal using conventional method; cost for intersystem handoff between subsystems and using LR-ING; cost for intersystem update procedure using LR-ING; cost for intersystem update procedure using conventional method; cost for intersystem registration using LR-ING; cost for intersystem registration using conventional method; total cost for location update using LR-ING; total cost for location update using conventional method; total costs for location update and paging procedures using LR-ING; total costs for location update and paging procedures using conventional method; paging delay in subsystem ; paging delay in subsystem . Let be the probability of intersystem mobility. Then, is given by [19] call duration

sojourn time a call arrives during

As a result, the total cost for paging a terminal using the LR-ING is expressed as (1)


In the same vein, the total cost for paging a terminal using the conventional method is expressed as

643

TABLE I LINK COSTS

(2) Moreover, the total costs for location update using the LR-ING involve registration, handoff, and update procedures in subsystem . It is expressed as (3) As a result, the total costs for location update and paging procedures using the LR-ING are given by

TABLE II DATABASE ACCESS COSTS

(4) On the other hand, the total cost for location update using the conventional method requires a supplementary update cost since there does not exist any equipment between the adjacent subsystems. In fact, when a subscriber in communication enters a new subsystem, the communication is temporarily suspended and restored once the subscriber registers in the new subsystem. As a result, the total update cost is given by (5) This enables us to evaluate the total costs for location update and paging procedures using the conventional method as follows:

whereas Figs. 1–3 enable us to, respectively, evaluate the cost for intersystem registration, the cost for intersystem updates, and the cost for paging a terminal by using the conventional method, as follows:

(6) To evaluate the performance of the proposed strategy, we combine (4) and (6) to obtain a cost ratio as follows: (7) CMR (call to mobility ratio) indicates the ratio where of the number of calls per unit of time to the number of changes of location areas per unit of time for a mobile terminal. Furthermore, to evaluate the costs of procedures defined in Section III and illustrated in Figs. 5–9, it is important to take into account the costs of accessing the databases, as well as the link cost illustrated in Fig. 4. In this context, Fig. 5 enables us by summing the costs of operations executed in to evaluate steps 2–6 as follows:

with as the number of adjacent subsystems. Moreover, to evaluate the intersystem paging delay, we have considered both one-step paging and sequential paging schemes, i.e., the mobile terminal may be located in one or more polling cycles [32]. In this context, when using the conventional approach, if a mobile terminal is searched, it will be paged in subsystem first. If this terminal is not found in subsystem , then it will be searched in subsystem . As a result, the intersystem paging delay using the conventional approach may be expressed as (8) where is the probability of intersystem mobility. However, when the LR-ING is used for intersystem paging, the mobile terminal is searched in only one subsystem. In this case, the intersystem paging delay is expressed as follows: (9)

Thus,

is given by

In the same vein, Figs. 6–9 enable us to, respectively, evaluate , , , and as follows:

where is the probability of intersystem mobility. In particular, when , the mobile terminal is located in one polling cycle [32]. In this case, we, respectively, obtain for (8) and (9): and . B. Numerical Results and Analysis For performance evaluation of the proposed architecture in terms of generated signaling costs, we consider three series of link costs (Table I) and three series of database access costs (Table II); whereas, the number of adjacent subsystems is ran. domly fixed to 5 and CMR

644

Fig. 10.


Fig. 11.

Cost ratio with link costs negligible and P = 0:8.

Fig. 12.

Cost ratio with database access costs negligible and various P .

Fig. 13.

Cost ratio with link costs negligible and various P .

Cost ratio with database access costs negligible and P = 0:3.

We first consider the case where the link costs are more sig; nificant than database access costs, i.e., whereas, and are given in Table I. Fig. 10 illustrates the behavior of the cost ratio given by (7) for . For the chosen values of link costs, the ratio is always smaller than one, which means that the proposed architecture transmits fewer signaling messages than the conventional architecture. The decreasing shape of the curves comes from the fact that, for small values of CMR, the users change location areas more often than they receive calls. Since the proposed architecture involves the LR-ING, which constitutes a supplementary link, it generates more signaling messages for such a situation. On the other hand, for low-speed users, the improvement in the signaling cost is about 45% for an intersystem mobility probability of 0.3. We have also considered the situation where the database access costs are more significant than the link costs, i.e., ; whereas, , , and are given in Table II. Fig. 11 illustrates the behavior of the ratio given by (7) for . For the given link costs, the proposed architecture is particularly efficient for series and , where the cost of accessing the LR-ING is less than or equal to the cost of accessing the HLR. More specifically, for series of Table II and high values of CMR (i.e., the users change location areas less often than they receive calls), the proposed architecture improves the conventional model by about 50%. However, when , the proposed architecture becomes less efficient than the conventional method since each location update is executed at the LR-ING, which generates more signaling traffic. Fig. 12 illustrates the behavior of the ratio given by (7) for various probabilities of intersystem mobility while considering that the link costs is more significant than database access costs, i.e., , and , (series 2 of Table I). Here, we have chosen to signify that the VLR manages all subscribers under its coverage; whereas, the LR-ING only manages the subscribers having changed subsystems. We realize that for the chosen link costs the performance of the proposed architecture increases in function of the probability of intersystem mobility. The best case corresponds to an improvement of the conventional model by about 60%.

Fig. 13 illustrates the behavior of the cost ratio given by (7) for various probabilities of intersystem mobility while considering that the database access costs are more significant than the link costs, i.e., , and , ,


645

with the HLR via the MSC. In this context, the total cost for location update using the BLR protocol is obtained by summing the costs of all operations executed during the same procedure of intersystem handoff, paging, or update. As a result, is given as (10) Furthermore, the cost for paging a terminal using the BLR protocol is calculated as

(11) Thus, the total costs for location update and paging procedures using the BLR protocol are given by (12) Fig. 14.

Ratio of intersystem paging delay.

(series of Table II). This supposes that the VLR manages fewer subscribers than the LR-ING and that the LR-ING man. ages fewer subscribers than the HLR, i.e., For the chosen link costs, the performance of the proposed architecture increases in function of the probability of intersystem mobility and for small values of CMR. In fact, when mobile users are moving from one subsystem to another, all update operations are executed at the LR-ING, avoiding the access to the HLR. In general, the proposed architecture improves the conventional model by about 60%. The performance of the proposed approach has also been , where and are, reevaluated by defining the ratio spectively, given by (8) and (9). Such a ratio quantifies the improvement of an intersystem paging delay obtained when using the proposed approach. Fig. 14 illustrates the behavior of this ratio in function of . We realize that the proposed architecture is capable of reducing the intersystem paging delay regardless of the paging scheme used in each subsystem. More specifically, if the mobile terminal is located in one polling cycle, i.e., , the proposed architecture improves the intersystem paging delay obtained from the conventional model by a maximum of 50%. This improvement may reach 65% when a sequential paging scheme is used in subsystem , or 30% when a sequential paging scheme is used in subsystem . C. Comparison With BLR Protocol In this section, we compare results obtained from the proposed architecture with those obtained from the BLR protocol (presented in Section II) [32]. In particular, we will estimate and compare the intersystem update and paging costs obtained by each procedure, which will enable us to conclude on the performance of each one. 1) Cost Evaluation: For comparison purposes, we define the following parameters for the BLR protocol: transmission cost from MSC/VLR to BLR; access cost to the BLR. These costs have the same characteristics as those defined in Section III, with the difference that the BLR can communicate

Moreover, in the context of the proposed architecture, only for location update changes, since we only the total cost consider the intersystem handoff process. We thus have

Then, the total costs for location update and paging procedures using the LR-ING are given by (13) For performance comparison, we combine (12) and (13) to define a cost ratio as follows: (14) where and are, respectively, given by (12) and (13). 2) Parameter Definition: For comparison purposes, the link costs are defined in Table III; whereas, the database access costs are defined in Table IV. Since is a parameter common to both architectures, it remains the same for the performance comparison. On the other hand, has been chosen to be less than, equal to, or greater than . For the database access costs, we consider that the VLR only contains the user information in its LA, which is always smaller than the LA controlled by the HLR, the LR-ING, or the BLR. As a result, is equal to one; whereas, all other cost parameters are greater than one. On the other hand, since the HLR serves more subscribers than the LR-ING or the BLR, we assume that and . 3) Results Analysis: This section compares the performance of the proposed architecture with the BLR architecture. This analysis is based on (14), while using link costs defined in Table III and database access costs defined in Table IV. We first consider the case where the link costs are more significant than database access costs, i.e., ; whereas, , , and are given in Table III. Fig. 15 illustrates the behavior of the ratio given by (14) for and CMR . We realize that when the BLR protocol gives better results due to the fact that, for the proposed architecture, all signaling messages sent from an MSC to the LR-ING must transit via the HLR, which increases the link costs. On the other hand, when , the proposed architecture transmits fewer signaling messages than the BLR

646


TABLE III LINK COSTS FOR COMPARISON WITH BLR PROTOCOL

TABLE IV DATABASE ACCESS COSTS FOR COMPARISON WITH BLR PROTOCOL

Fig. 15. Comparison with BLR protocol (database access costs negligible and P = 0:3).

architecture. Furthermore, when , the proposed architecture enables us to reduce the signaling costs obtained from the BLR protocol by about 40%. The decreasing shape of the curves comes from the fact that, for high values of CMR, the scenario for paging a terminal becomes more efficient using the proposed architecture. On the other hand, for low-speed users, the improvement is about 45% of the total cost for a probability of intersystem mobility of 0.3. We now consider the situation where the database access costs are more significant than the link costs, i.e., ; whereas, , , , and are given in Table IV. Fig. 16 illustrates the behavior of the ratio given by (14) for and CMR . For the chosen values of database access costs, we realize that both architectures give similar results in terms of the number of generated signaling messages. The main difference resides in the interconnection of the added equipment: the LR-ING is connected with the HLR of each subsystem, whereas the BLR is connected with all MSC/VLR in each subsystem, which is

Fig. 16.

Comparison with BLR protocol (link costs negligible and P = 0:3).

Fig. 17.

Comparison with BLR protocol for various CMR.

financially more expensive for an operator who intends to adopt such an architecture. Finally, Fig. 17 illustrates the behavior of the ratio given by (14) in function of for various values of CMR . The other parameters are chosen as follows: , , , , , , and . For such parameters, we realize that the more the CMR increases, the more the proposed architecture improves results obtained from the BLR protocol. The increasing shape of the curves comes from the fact that when increases, the quantity of signaling messages generated by the proposed architecture becomes more important. In fact, for the BLR architecture, each MSC/VLR has a direct access to the BLR, which is not the case of the proposed architecture. Furthermore, when a request for location update is sent to the BLR and the user has not changed subsystem yet, the BLR does not know exactly in which location area to search. Consequently, it sends a signaling message to the HLR so that the latter can find the mobile user. This explains why the proposed architecture gives better results when tends to zero. In general, the improvement reaches 30% approximately.


V. CONCLUSION In this paper, we have proposed an architecture and new schemes for intersystem registration, updating, and search scenarios in the context of global mobility management. The proposed architecture essentially consists of adding at the boundary location area of two adjacent subsystems a specialized piece of equipment called the LR-ING. The latter is connected to the HLR of each subsystem and maintains roaming information on mobile users moving between different subsystems. This enables us to minimize the number of operations executed by the databases in the context of global mobility management. Numerical results have shown that the proposed scheme enables us to significantly reduce the signaling cost generated by the databases, as well as the intersystem paging delay. More specifically, the reduction cost reaches about 50% when comparing the proposed architecture with the conventional architecture and about 20% when comparing the same architecture with the BLR protocol. Further work should be oriented toward developing new strategies which decorrelate location update and search procedures and which take into account the user classes in terms of mobility and type of generated traffic. Then, it would be interesting to implement and test such strategies with real-time simulation software like or NS-2. OPNET ACKNOWLEDGMENT The authors would like to thank the anonymous reviewers for their constructive remarks and suggestions. REFERENCES [1] L. P. Araujo and J. R. B. de Marca, “A comparative analysis of paging and location update strategies for PCS networks,” in IEEE Int. Conf. Commun., Atlanta, GA, 1998, pp. 1395–1399. [2] A. Bar-Noy and I. Kessler, “Tracking mobile users in wireless communications networks,” in Proc. 12th Annu. Joint Conf. IEEE Comput. Commun. Soc., vol. 3, San Francisco, CA, 1993, pp. 1232–1239. [3] F. V. Baumann and I. G. Niemegeers, “An evaluation of location management procedures,” in Proc. IEEE ICUPC, San Diego, CA, Sep. 1994, pp. 359–364. [4] R. Beaubrun, “Gestion de la mobilité et ingénierie de trafic en conception de réseaux mobiles de troisième génération,” Ph.D. dissertation, Dept. Elec. Eng., École Polytechnique de Montréal, Montreal, QC, Canada, 2002. [5] A. Bertrand, “Jambala mobility gateway-convergence and inter-system roaming,” Ericsson Rev., vol. 76, pp. 86–93, 1999. [6] J. M. Brazio and N. J. S. Silva, “Performance evaluation of a multilayer location update method,” in Proc. IEEE Veh. Technol. Conf., vol. 1, Atlanta, GA, 1996, pp. 96–100. [7] T. X. Brown and S. Mohan, “Mobility management for personal communications systems,” IEEE Trans. Veh. Technol., vol. 46, no. 2, pp. 269–278, May 1997. [8] K. Buchanan, R. Fudge, D. McFarlane, T. Phillips, A. Sasaki, and H. Xia, “IMT-2000: Service provider’s perspective,” IEEE Pers. Commun., vol. 4, no. 4, pp. 8–13, Aug. 1997. [9] “Digital cellular telecommunications system (Phase 2+) (GSM); Universal mobile telecommunications system (UMTS); Location management procedures,” ETSI, TS 123 012 V4.0.0 (2001-03), 3GPP TS 23.012 version 4.0.0 Release 4, 2001. [10] D. Gu and S. S. Rappaport, “A dynamic location tracking strategy for mobile communication systems,” in Proc. IEEE Veh. Technol. Conf., vol. 1, Ottawa, Canada, May 1998, pp. 259–263. [11] M. Hellebrandt and R. Mathar, “Location tracking of mobiles in cellular radio networks,” IEEE Trans. Veh. Technol., vol. 48, no. 5, pp. 1558–1562, Sep. 1999.

647

[12] J. S. M. Ho and I. F. Akyildiz, “Dynamic hierarchical database architecture for location management in PCS networks,” IEEE/ACM Trans. Netw., vol. 5, no. 5, pp. 646–660, Oct. 1997. [13] , “Mobile user location update and paging under delay constraint,” Wireless Networks, vol. 1, no. 4, pp. 413–425, 1995. [14] R. Jain, Y. B. Lin, C. Lo, and S. Mohan, “A caching strategy to reduce network impacts of PCS,” IEEE J. Sel. Areas Commun., vol. 12, pp. 1434–1444, 1994. [15] J. Jannink, D. Lam, N. Shivakumar, J. Widom, and D. C. Cox, “Efficient and flexible location management techniques for wireless communication systems,” in Proc. 2nd ACM Int. Conf. Mobile Comput. Netw., Nov. 1996, pp. 38–49. [16] K. K. Leung and Y. Levy, “Global mobility management by replicated database in personal communication networks,” IEEE J. Sel. Areas Commun., vol. 15, no. 10, pp. 1582–1596, Oct. 1997. [17] J. Li, Y. Pan, and X. Jia, “Analysis of dynamic location management for PCS networks,” IEEE Trans. Veh. Technol., vol. 51, no. 5, pp. 1109–1119, Sep. 2002. [18] Y. B. Lin and I. Chlamtac, “Heterogeneous personal communications services: Integration of PCS systems,” IEEE Commun. Mag., vol. 34, no. 9, pp. 106–113, Sep. 1996. [19] J. McNair, I. F. Akyildiz, and M. D. Bender, “An inter-system handoff technique for the IMT-2000 system,” in Proc. 19th Annu. Joint Conf. IEEE Comput. Commun. Soc., Tel Aviv, Israel, 2000, pp. 208–216. [20] A. Noerpel and Y. B. Lin, “Handover management for a PCS network,” IEEE Pers. Commun., vol. 4, no. 6, pp. 18–24, Dec. 1997. [21] R. Pandya, D. Grillo, E. Lycksell, P. Mieybegue, H. Okinaka, and M. Yabusaki, “IMT-2000 standard: Networks aspect,” IEEE Pers. Commun., vol. 4, no. 4, pp. 20–29, Aug. 1997. [22] S. Pierre, Réseaux et Systèmes Informatiques Mobiles: Fondements, Architectures et Applications. Montreal, QC, Canada: Presses Internationales Polytechnique, 2003. [23] J. F. Rizzo and N. R. Sollenberger, “Multiplier wireless access,” IEEE Pers. Commun., vol. 2, no. 3, pp. 18–30, Jun. 1995. [24] R. D. Ryan, “Roaming between heterogeneous 3rd generation wireless networks,” in Proc. IEEE Wireless Commun. Netw. Conf., vol. 1, New Orleans, LA, Sep. 1999, pp. 321–323. [25] H. Safa, S. Pierre, and J. Conan, “A built-in memory model for reducing location update costs in mobile wireless network,” Comput. Commun., vol. 25, no. 14, pp. 1343–1353, Sep. 2002. , “An efficient location management scheme for PCS networks,” [26] Comput. Commun., vol. 24, no. 14, pp. 1355–1369, Aug. 2001. [27] S. Tabbane, “An alternative strategy for location tracking,” IEEE J. Sel. Areas Commun., vol. 13, no. 6, pp. 880–892, Jun. 1995. [28] Q. Tian and D. C. Cox, “Location management in a heterogenous network environment,” in Proc. IEEE Wireless Commun. Netw. Conf., Chicago, IL, Sep. 2000, pp. 753–758. [29] K. Ushiki and M. Fukazawa, “A new handover method for next generation mobile communication systems,” in Proc. IEEE GLOBECOM, Sydney, Australia, Nov. 1998, pp. 1118–1123. [30] W. Wang and I. F. Akyildiz, “A predictive paging scheme for IMT-2000 systems,” in Proc. IEEE Veh. Technol. Conf.-Fall , vol. 3, Boston, MA, Sep. 2000, pp. 1424–1430. , “Intersystem location update and paging schemes for multitier [31] wireless networks,” in Proc. MobiCom Conf., Boston, MA, Aug. 2000, pp. 99–109. [32] W. Wang, “Location management techniques for next generation wireless systems,” Ph.D. dissertation, School Electrical Comput. Eng., Georgia Inst. Technol., Atlanta, GA, 2002. [33] J. Xie and I. F. Akyildiz, “A novel distributed dynamic location management scheme for minimizing signaling costs in mobile IP,” IEEE Trans. Mobile Comput., vol. 1, no. 3, pp. 163–175, Jul.–Sep. 2002. [34] H. Xie, S. Tabbane, and D. J. Goodman, “Dynamic location area management and performance analysis,” in Proc. IEEE Veh. Technol. Conf., Secaucus, NJ, May 1993, pp. 536–539.

Abdoul Djalil Assouma was born in Cotonou, Republic of Benin, on July 1, 1979. He received the B.Eng. degree in electrical engineering and the M.Sc.A. degree in computer engineering from École Polytechnique de Montréal, Montréal, QC, Canada, in 2002 and 2004, respectively. His research interests include mobility and quality-of-service problems in the next-generation wireless systems. He is currently working on telecommunication network management in Montreal.

648


Ronald Beaubrun (M’02) received the B.Eng., M.Sc.A., and Ph.D. degrees in electrical engineering from École Polytechnique de Montréal, Montréal, QC, Canada, in 1994, 1996, and 2002, respectively. From 1994 to 1999, he worked as a Research Assistant at the LICEF Research Centre, Montréal, and Ericsson Canada, where he contributed to many projects related to multimedia telecommunications systems, virtual campus, virtual laboratories, and reliability of 3G wireless networks. He is currently an Assistant Professor at Université Laval, Laval, QC, Canada, where he teaches computer networks and mobile communications in the Department of Computer Science and Software Engineering. His research interests include topics related to the next-generation wireless networks planning, such as radio coverage, architecture, global roaming, resource management, traffic modeling, as well as value-added services and applications.

Samuel Pierre (SM’97) received the B.Eng. degree in civil engineering from École Polytechnique de Montréal, QC, Canada, in 1981, the B.Sc. and M.Sc. degrees in mathematics and computer science from the UQAM, Montréal, in 1984 and 1985, the M.Sc. degree in economics from the Université de Montréal, in 1987, and the Ph.D. degree in electrical engineering from École Polytechnique de Montréal, in 1991. From 1987 to 1998, he was a Professor at the Université du Québec à TroisRivières, prior to joining the Télé-Université of Québec, an Adjunct Professor at Université Laval, Ste-Foy, Québec, an Invited Professor at the Swiss Federal Institute of Technology, Lausanne, Switzerland, and the Université Paris 7, France. He is currently a Professor of Computer Engineering at École Polytechnique de Montréal, where he is Director of the Mobile Computing and Networking Research Laboratory (LARIM), Chairholder of the NSERC—Ericsson Chair in Next Generations Mobile Networking Systems, and Director of the Mobile Computing and Networking Research Group (GRIM). His research interests include wireline and wireless networks, mobile computing, artificial intelligence, and telelearning. Dr. Pierre is a member of the Association for Computing Machinery (ACM). He is a Regional Editor of the Journal of Computer Science, an Associate Editor of IEEE COMMUNICATIONS LETTERS, and the IEEE Canadian Review, and serves on the Editorial Board of Telematics and Informatics, edited by Elsevier Science. He has received many distinctions such as the Prix Poly 1873 for excellence in teaching, in 2001, and Fellow of the Engineering Institute of Canada, in 2003, among others.