TPRLM: Time-based Probabilistic Relational Location Management Scheme for Wireless Cellular Networks Sourav Saha1, Mainak Mukherjee2, Avik Ranjan Paul2 and Sarmistha Neogy3 1

Dept. of Information Technology, Bengal Institute of Technology, Kolkata, India, [email protected] 2 Dept. of Electronics & Communication Engg., Bengal Institute of Technology, Kolkata, India [email protected], [email protected] 3 Dept. of Computer Sc. & Engg., Jadavpur University, Kolkata, India, [email protected] Abstract

Locating mobile nodes and delivery of inward calls to that node in wireless cellular networks, in spite of extreme mobility of the nodes, is a challenging issue. Fast and error-free delivery of packets to perfect destination with good performance in wireless environment is extremely difficult to attain. Several available location management schemes incur huge location update cost, paging cost and network bandwidth wastage in attempt to achieve it. They also offer significant location update delay and paging delay. Interactivity and great efficiency are also lacking. In this paper, the framework of a scheme TPRLM is proposed. This scheme combines relational levelbased coverage area management with organized Informing Cells and time-based probabilistic priority calculation based on inward call arrival rate of any node. The scheme tries to attain cost-effectiveness, reduced location update delay and paging delay. There is also scope for interactivity and improved performance without complexity.

I. INTRODUCTION “Working while moving” is an extremely essential requirement nowadays. But error-free delivery of packets to perfect destination in network consisting of mobile hosts is truly challenging. The problem of maintaining end-to-end consistent service for wireless networks increases with the increase in number of nodes in the network or boost of random mobility of nodes. The approximate locations of any node can be obtained through use of some technique. The movement of any node may be tracked also. The source node may ask for the approximate location of the destination node from the location server who contains updated information. The source node, at receipt of location of destination node can then start routing packets to it. Several location management strategies are available for wireless cellular networks. But the paging and location update cost is not too small. The network bandwidth is wasted also. The location update time and paging delay has scope to be further improved. The increase of interactivity is also very much required. The overall performance of location management has provision for supplementary enhancement without making the procedure extra ordinarily complex. In this paper the basic concept of location management along with several types of locating any node in wireless environment are discussed. The basic ideas of various scheme proposals related to location update, paging and mobility agent

based location management are critically analyzed to detect their advantages and future development scopes. Then we have proposed a scheme named TPRLM, where through invocation of time-based probabilistic relational location management, the paging and location update cost, network bandwidth wastage, location update and paging delay etc. are attempted to reduce. The interactivity and performance of location administration are tried to decrease here. The remaining part of the paper is organized as follows: Section II offers the basic ideas on location management in context of mobility handling. Section III presents a report of the survey of previous works in related areas for cellular networks and section IV describes the overall status of the already proposed schemes. Section V mentions the basic objectives of our approach. In section VI, the framework of an improved scheme TPRLM is proposed along with its basic concept, algorithm, illustration etc. Finally in section VII, we conclude with the future development scopes of our work. II. LOCATION MANAGEMENT: BASIC CONCEPTS Location Management is an interesting and immensely necessary service for superior performance of frequently moving nodes in wireless network. This enables any mobile node in wireless network to know the approximate location of any mobile node so that it can send the packets to the actual destination node in spite of its movement. In location management, the success of sensing the mobile node location, movement detection of that node etc. performs major role in delivery of packets to accurate destination. Hence location management basically consists of three steps: location registration (selecting location servers, authenticating and periodically updating location server database with approximate location information), location maintenance (maintenance of location database by cleansing it periodically for better performance and database consistency) and call delivery through location discovery (finding out the approximate location of the destination node through query made by the source node in the location server database and delivery of call to the destination node). Any sender mobile node can locate the destination mobile node in mainly two ways: through continuous location tracking or through location searching immediately before interaction. The methods are depicted in the figure 1 below.

Locating any Mobile Node

Continuous Location Tracking

Location Searching Immediately before Interaction

Figure 1: Techniques to locate any mobile node

arranged into several categories [1, 4, 5] that is reflected in figure 2 next. Location update schemes can be local or global. They may also have several subcategories. Local Location Update Scheme: Here any particular mobile user can choose the time and area for itself for initiation of location update process. It is not dependent on the location update areas of other mobile users. These strategies can be divided into two subcategories: static and dynamic. We shall Location Update Schemes

For continuous location tracking the mobile node or the mobility management system has to constantly sense locations of nodes in the network. The movement of any node is detected and maintained in some storage area for future use. For short-timed asynchronous communication where several source mobile hosts send packets frequently to any destination mobile node, the continuous location tracking method is suitable. In such a scenario the mobile nodes have to contribute to the location management system by specifying the frequently changing locations in order to keep the system informed or the mobile node itself must be responsible for continuous tracking of node location. Depending on the change of locations of mobile nodes location updates are made here. Interaction between source node and destination node may commence on successful detection of the location of the destination mobile node at that time. In case of location searching immediately before interaction, the location of any mobile node is not constantly tracked. This method of locating any mobile node is appropriate for longterm peer-to-peer communication. Paging process can be used here to find out the approximate location of any mobile node. The network coverage area is divided into several location areas. When a call is to be processed the destination mobile node is found out through paging. The mobility agents may be used in sensing the location of any mobile node directly before any communication. III. LOCATION MANAGEMENT IN CELLULAR NETWORK: A CONCISE REVIEW OF THE PREVIOUS RELATED WORKS There are various proposals related to managing locations of mobile nodes in cellular environment. We have studied the basic concepts of such schemes and presented a critical analysis of our survey in this paper. Any mobile node can be located in cellular networks basically in two ways, as have been specified in section II. Here we have reviewed the areas of location update, mobility agent and paging in separate subsections. A. Location Update Schemes This is a procedure through which any mobile node makes its latest location known to the system. Here one mobile node can help the network to update its database containing location information of the mobile nodes by messages. The updating of location information content of that database helps the system to deliver any call to the destined mobile node in spite of its immense mobility. The location update schemes can be

Local

Static

Global

Dynamic

Static

Location Area-based

Time -based

Movement -based

Path Report -based

Topology -based

Per-user based

Profile -based

Dynamic

Reporting Center-based

Loadsensitive Distance -based

Figure 2: Several location update schemes in organized manner

call them static local and dynamic local location update schemes respectively for better characterization. They are discussed below. • Static Local Location Update Scheme: In these schemes the network is assumed to consist of explicit cells or areas where location updates occur when any new mobile node visits that area. The generation of location updates are of static nature, since the cells or areas for generating location updates are preset and do not change depending on the movement patterns of the mobile nodes. In individualized location area (LA) update scheme [4, 8], the location update takes place when any mobile node comes into a pre-assigned set of location areas. • Dynamic Local Location Update Scheme: In this type of location update schemes, the updating of location information occurs based on the movements of the mobile nodes instead of any pre-defined set of location areas. We now discuss several dynamic local location update schemes below: o Time-based Strategy: Several time-based strategies [4, 5, 6, 9, 10] are available where consecutive location updates occur using fixed or variable time intervals (the interval is calculated differently in different schemes. The most straightforward proposal in this context is the time-based scheme [6], where location updates are done by any mobile station periodically every T amount of time. But here a lot of unnecessary updates are to be performed independent of movement of the nodes.

o

o

o

o

Some enhancements are found in the timer-based scheme proposal [9], where the time interval between two consecutive location updates is not a fixed time interval like the previous approach. Here an attempt is made to reduce the average cost of updating location information through calculating maximum waiting time between two successive location update operations. But the improvement achieved is not very much than the other. In time-based, dynamic mobile user location update scheme [10], the decision about the location update time is taken by the mobile node dynamically based on the node movement pattern in the network and the frequency of inward call. A location update occurs if the paging cost to serve any inward call to it is greater than the cost of location updating. This time-based dynamic location update scheme shows improvements in performance compared to the fixed time-based location update scheme. But from overall point of view of location management performance, this scheme fails to put up to requirement. We discuss the following techniques together below. Path Report-based Strategy: LeZi-Update scheme [34] falls into this category since here the movement path of any mobile node is maintained based on reports given by the nodes. Path-report is delivered if the movement path is not already stored. During inward call delivery to that node the probable path is estimated using previously stored movement paths of that node. Location update cost and paging cost is small here. But the probable path estimation requires some more advancement. Movement-based Scheme: There are several movement -based [4, 6, 46] schemes where location updates are made based on mobile node movement. In classical movement-based scheme [6], any mobile node increments a counter value on crossing of cell border and location is updated and counter value is reset on attaining of the counter value any predefined value M. The scheme is truly very easy. But the paging cost is needed to be reduced, since for inward call delivery to any node here, the system has to page every cell within a space of M from the lastly updated cell. In an enhanced movement-based scheme selective paging process is proposed for decreasing the total location management cost of system. The improvement attained through is small. Per-user based Strategy: In this type of approach [47] the location updates are performed dynamically based on the movement types or some other information of individual mobile nodes. The performance of this easy scheme is better than static ones. Topology-based Strategy: In this scheme [33] the present location of any mobile node can be determined using the previous and present location of that node. The location is updated only if the previous location is varying from present location. Here location update

cost is little. But this scheme is too straightforward for dealing with practical location management of nowadays. o Load-sensitive Strategy: In [35], the deployment of non-utilized channel bandwidth is proposed for generating queries about the node location. This scheme truly decreases the location management cost. But in frequently moving environment this scheme does not perform up to mark. o Profile-based Strategy: There are several mobile node user profile-based schemes [4, 31, 32] available as surveyed below: In one profile-based strategy [31], two approaches, based on maintenance of user movement pattern as mobile node user profile, are offered. In first approach, the long-standing information related to location areas and the user location probability in that area is saved for further need. If the mobile node does not visit any new location area (i.e. not stored by system) in any particular time, the location updating is not performed. On appearance of inward call destined for the mobile node, the location area with maximum node location probability for any particular time, as saved by the system, is paged. If not found, the stored location area with next highest node location probability is paged. The first approach definitely diminishes the cost of location updating through storage of location information. The second approach keeps track of longstanding as well as little or medium-standing actions also to reduce the location management cost, which necessitates more storage space. In another profile-based strategy [32], the parameters used for defining the profile of any mobile node are the total occurrence of cell switches performed by any node and the average amount of time of node attendance in every cell. The maintained profile helps the system to dynamically produce the location areas with estimation of the probability for the presence of any mobile node there. The location is updated only if the node visits any new area that is not already saved. On appearance of inward call destined for any mobile node the paging initiates from the saved area where node attendance time is maximum. If found, others are not paged; else area with second highest node presence time is paged and so on. The location update cost and paging cost here are not big. But if any node switches cells randomly with short span of presence in any cell, the total location management cost increases. o Distance-based: Various distance-based schemes [1, 4, 5, 6, 12, 13, 14, 15] exist. We analyze them below: The simplest one is the distance-based location update scheme [6], where location is updated when the distance from the previous updating location to recent position attains a preset distance threshold D. It shows improved performance as compared to several schemes through limiting the search area etc. But this scheme is

very simple in terms of coexistence with several mobility patterns of nowadays. Another distance-based location update scheme [12] employs selective paging in addition to simple distancebased concept with D. Here the area with radius D is partitioned into sub-areas where polling is done one-byone to find any new mobile node. This scheme shrinks the average total cost of location update and paging as compared to the previous one. However it complicates the total management. In look-ahead strategy for distance-based location tracking [13], optimal future cell is searched so that it may be notified during location update. Here the frequency of updating location information is considerably decreased. And it improves the location management system performance marginally. The predictive distance-based mobility management scheme [14], attempts to predict the future location of any mobile node on the basis of its most recent position and velocity. It enhances the performance of the location management as compared to simplistic distance-based approach. But the prediction is approximate. In [15], distances are represented using coordinate systems to improve cell addressing process for location management. But it helps in methodical calculation process only. Global Location Update Scheme: Here updating is carried out in identical set of cells by all the mobile node users. It is not done for any single mobile user. It can be either static or dynamic. • Static Global Location Management Scheme: Here location updating occurs for predefined set of cells by all the mobile nodes. It can be of two flavors as mentioned below: o Location Area-based Scheme: There are various location area-based approaches [1, 4, 5, 6, 16, 17]. In fundamental location area-based scheme [6], the network coverage range is divided into location areas (LAs) in a way to decrease the costs for location update and paging. Any mobile node updates location information in case of switching from one LA to another. But for mobile node that frequently switches from one LA to another, staying very short time in any particular LA, the overhead for location update therefore increases significantly. The selective location update strategy [16] attempts to solve the aforesaid problem by updating location information only on entering some pre-determined set of cells based on the movement patterns of the mobile nodes. But for high user location probability and high incoming call appearance rate the performance may be marginally better over others. In the context of location area-based approaches we like to mention another a1gorithm, named Two Location Algorithm (TLA) [17], which attempts to decrease the location update cost through allowing any mobile node

to uphold only two most currently stayed location areas. If the movement of the mobile node is limited to currently visited cells most of the times, this scheme offers very good result. o Reporting Center-based Scheme: This strategy [4, 30] employs some chosen cells from the system coverage area as reporting centers which are used by the mobile nodes as updating points of location information. Any mobile node, entering any newer reporting center area, updates its location information. The reporting centerbased concept is balanced and organized. But the paging here is costly enough because all the cells near the currently updated reporting center are paged in order to locate any mobile node before delivery of any inward call destined for it. • Dynamic Global Location Management Scheme: In this category all mobile nodes update location information in identical set of cells on the basis of the mobility patterns of the nodes. One example of this type of strategy is timevarying location area strategy [4, 18], where global location updates are performed with variable interval. It performs better than fixed time-based location update approach. But the improvement achieved is little. B.

Mobility Agent–based Schemes

Mobility agents (namely home agent and foreign agent) [19, 20, 21, 22] aid in movement of any mobile node from one network (home network) to another (foreign network) and still maintain connection with the previous network. Several protocols that use mobility agent-enabled location management are surveyed briefly below. Mobile IP, Mobile IPv6 etc: In MIP [19, 20, 21, 22], the CoA (Care of Address) helps in locating any mobile node moved from its home network to any foreign network. CoA may be generated using foreign agent (FA) in foreign network. The home agent (HA) can maintain and update the MIP registration table. MIPv6 [3, 22, 36] is advanced version of MIP to work with IPv6. The fundamental concept is still present nowadays in spite of several advancements. But decreased power consumption, advanced QoS etc. are required. CIP: In Cellular IP[3, 22, 37, 38] different access points in various cells are connected through Gateway Router and offers fully static nodes, repeatedly moving nodes and infrequently moving nodes very good location management performance. But packet loss rate is not little here. HMIP and HMIPv6: In Hierarchical MIP [3, 22, 39] the hierarchical arrangement of FAs enables the Gateway Foreign Agent at root to track the movement of the mobile nodes at leaves of the hierarchy. The paging facility and hierarchical structure aid in improvement of performance. But it suffers from scalability problem. IDMP (Intra-Domain Mobility Management Protocol): In this lightweight scheme [22, 40, 41] node location is managed in spite of their mobility through using several

Subnet Agents in several subnets connected to the root (Mobility Agent). Here use of modular approach, paging and two dynamically autoconfigured CoAs reduces the power consumption and packet loss. But it needs to be a bit more secured. TeleMIP: Telecommunication Enhanced MIP [22, 42, 43] employs a two-tier hierarchical structure with several GFAs at root for mobility management. Scalability is a very good advantage for it. But as it makes use of MIP for global mobility management, the drawbacks of MIP distresses the efficiency of it. TLMM and TLMIPv6: Three Level Mobility Model [22, 44] is a three-layer scheme using Global Mobility Agents at root level for internetworking, Mobility Agents at middle layer for managing intra-domain mobility. Three Layer Mobile IPv6 scheme [22, 45] is enhanced version of TLMM for working with IPv6. Employment of three CoAs for three layers helps in better location management. The signaling load here is less. C.

Paging Schemes

The location of any mobile node can be found out before interacting with that node through paging. Paging aids in call delivery process by locating any mobile node, in spite of its mobility between cells, on the basis of location information present in the system. Several paging schemes are present as studied below: Sequential Paging Strategy: In this strategy [23], the location area is divided into sequentially numbered cells depending on the location probability of mobile node in the cells. This strategy is well organized, but the paging cost here is more. Reverse Paging Scheme: This scheme [5, 24] attempts to reduce the cost of paging through dividing the service area into several partitions and assuming the probable location of any mobile node from another one is little. But this assumption is not always practical. Semi-reverse Paging Scheme: In this scheme [5, 48] the coverage area of mobile nodes is divided into several paging areas on the basis of non-increasing sequence of probability of mobile node locations. The use of advanced concept of location management improves the performance. But paging delay can be reduced further. Selective Paging: In this scheme [1, 28, 29] the coverage area is divided into several areas based on the mobility pattern of the node and the inward call arrival rate to it. In order to locate any mobile node there one area is paged for some fixed amount of time, instead of all the areas at a time. If the node is found, other areas are not searched at all. If not found, another area is paged. The paging cost here is definitely less. But the amount of time required to locate any mobile node in average practical situation here is not little. Cluster Paging Strategy: The coverage area in this scheme [1, 10] is divided into concentric rings each consisting of equal sized hexagonal cells of identical mobile node location probability (for the cells within any particular ring

the probability is identical). The location probability of mobile nodes in the inner ring cells are greater than equal to that of outer ring cells and the inner rings contain lesser number of cells than outer rings. Any particular mobile node is searched here starting from inner ring cells and continuing toward outer ring cells. The organized management of the coverage area absolutely advances the location management process here. But in real life scenario it is very difficult to find many mobile nodes with such ordered node location probability. Uniform Paging Scheme: It is a very simple scheme [5, 26] where the number of divisions of every paging area is roughly the same. This reduces the complexities of the management process, but cannot decrease paging delay. Intelligent Paging Strategy: This scheme [5, 27] employs multi-step method for locating any moving mobile node. On appearance of any inward call, the destined mobile node is located through mapping the paging area within the location area on the basis of moving speed of node, node state, rate of inward call appearance to that node and its mobility pattern. Here the mapping process aids a lot in finding the exact paging area of any mobile node within the coverage area. Signaling overhead is not huge here. But repetitive paging failures can result in increase in paging cost here. IV. CRITICAL ANALYSIS OF THE OVERALL STATUS OF LOCATION MANAGEMENT SCHEMES IN CONTEXT OF MOBILITY HANDLING After studying various location management schemes for cellular networks either continuous location tracking or searching location immediately before interaction, we present a brief summary of our survey below. The schemes attempt to decrease the signaling traffic overload, location update or paging cost and delay in communication. Efficient coverage area partitioning, improvement of paging techniques, reduction of the number of location updates, improvement of system performance have been the general endeavor of several proposals. But the issues of reporting area identification, location-tracking delay, location updating and paging cost etc. have to be developed further for better performances. V. BASIC OBJECTIVES OF OUR APPROACH The basic goals of our proposed scheme are as follows: a. Reducing the paging cost. b. Decreasing the wastage of network bandwidth. c. Lessening of location update delay and paging delay. d. Increasing interactivity. e. Improvement of performance of location management with the minimum possible overhead. VI. TPRLM - TIME BASED PROBABILISTIC RELATIONAL LOCATION MANAGEMENT A. Introduction

The various location management schemes as studied in section III focus on different areas for improvement. The location update and paging delay, channel bandwidth wastage, paging cost etc. are required to be minimized for betterment. Again, performance and interactivity enhancement are of utmost demand. Here we present the framework of a location management scheme named TPRLM that tries to reduce the delay for location update and paging. We have also tried to lessen network bandwidth wastage and paging cost. This scheme also intends to make the method dynamically interactive. The overall aim of this scheme is to improve the performance of location management process through cost effectiveness, network bandwidth utilization, interactivity and faster location update and paging procedure without incurring too much overhead. B. The Concept of Relational Location Management In our scheme the location of a mobile node is tracked through invocation of related level-based coverage area management concept. We consider the latest informed cell to be the home location of that mobile node. When a call arrives the tracking begins from the last informed visited cell (i.e. the present home location). Paging initializes query for the end point location and all coverage area cells send beacons so that the intended mobile station responds that would ultimately result in an update to the location register. When the search in home location fails, 1 st Level (i.e. the immediate neighbors of the cell) is looked for. The tracking may continue up to n th Level of cells if the desired node is not found. The structures thus formed can be termed as composite cells. In figure 3 below, we have shown it up to 3rd level. So the home location searches its neighbors and then those cells that share a common boundary with the neighbours, which in turn carries on the same procedure until the mobile station is found and the location is updated. The search stops when the mobile station first responds to the paging and identifies itself by responding to the nearest cell. Home Location 1st Level Cell 2nd Level Cell 3rd Level Cell

The total number of cells found in an n Level linear structure can be obtained from below, where n is the order of the Level. j = 1 + 6 ∑ n , where j is the maximum Level i =1 =1+6(1+2+3+4+5+....n ) =1+6

( n )( n + 1) 2

The above stated technique reduces the paging cost and total location management cost. The delay of location update and paging also diminishes through this approach. So, the performance will improve. C. The Concept of Time Based Probabilistic Location Management We employ time-based probabilistic technique, detailed below, for locating mobile nodes in our scheme. It can be seen that as the mobile station moves far away from the recent home location, the cost of paging increases as the mobile station moves far away from the last tracked home location. If the number of level increases the paging cost increases (paging cost α n) . The call arrival probability can play a very important role in this context. If the call arrival time is known to the called mobile station in advance, the mobile station can update its location just before the call arrival time. In this way, both costs of location update and paging can be kept optimal. So the call arrival probability for any mobile node may be estimated. Calculation of the Relative Call Arrival Rate It is fact that all mobile nodes in any area do not receive the same amount of calls. It is assumed that there are mobile nodes that receive more than a hundred calls per day while others receive around ten calls per day or even lesser than that. So in order to reduce the paging cost the frequency of paging the former mobile nodes would be greater than the later. Let the mobile stations be distinguished by their call arrival rates in the previous month. So the first step is to group the mobile station according to their call arrival rate (say β). This may be done by calculating the call rate of the previous month (say a month of 30 days). Let the number of arrived call be n1 for a mobile node say m1 and similarly call arrived be n2 for a mobile node say m2. We assume that n2 is higher than n1. The Group

Figure 3: Relational level-based coverage area division forming composite cells

The number of paging increases as the mobile station travels far and wide. The number of cells in any level may be found out from the series shown below, where the leftmost element represents the number of cells for home location, the second element from left is number of cells for 1 st level and so on. th 1+6+12+18+24+... upto n term

No of Calls per month (n)

Call Arrival Rate β =

β

n 30 × 24 × (60)

2

Priority factor

Group I

>>50000

Group II

49999-25000

0.009645 >> β <<0.01928

2

Group III

24999-15000

0.009654 >> β << 0.00578

3

Group IV

14999-5000

0.00578 >> β << 0.00192

4

Group V

<<5000

β

>> 0.01929

<<0.00192

Table 1:Calculation of the call arrival rate

β

1

5

and setting priority respectively

inward call arrival rate (number of calls/sec) for a month for m1 is: n1 / (30*24*3600) and for m2 is: n2/ (30*24*3600). It can be easily concluded that first one is lesser than the second one. Thus the mobile nodes may be distinguished into groups according to different rates of incoming calls. Assigning priority to more probable call arrivals Now a priority (named as priority factor) is given to all mobile nodes according to the group in which they are classified based on the call arrival probability. The group of mobile nodes whose call arrival rate is higher than the rest is given higher priority i.e. lower priority factor value (priority factor value 1 means highest priority). Therefore more probable mobile node would be tracked sooner than the less probable mobile nodes. In table 1 before, we have shown the calculation of call arrival rate (β) and setting of priority to them. In table 2 below, we depict the variation of time interval of paging based on priority using 5 different priority levels. Priority factor 1 2 3 4 5 Table 2

paging all the cells. In these diagrams, different shades are used that denote 1st, 2nd, 3rd and 4th levels consecutively. Figure 4 shows how the mobile station is tracked from its home location to the cell where the mobile node visits. The process took two Levels to track the mobile station. After tracking the mobile node’s visiting cell, it becomes the mobile node’s current home location. In this case, instead of paging all the cells, a small group of cells are only paged that reveals the mobile node’s current location which is correspondingly updated. The map of the network that is shown below shows how the mobile station is tracked in its movements responding to its visit in seven locations. The following figure gives emphasis and reveals the TPRLM realization a bit more clearly. When the above system is realized for multiple mobile nodes then the time based probabilistic model is used in conjunction with the current model which sets a priority for all groups.

Tracking interval (Sec) 0.1 1 10 100 1000

Division of time interval of paging in accordance to priority Figure 5

Figure 6

D. Time Based Probabilistic & Relational Location Management: A Model Case Study Now we discuss the total scheme through a case study. We consider a region of cellular network as sample where seven movements of a mobile node is being attempted to track. Here the cells in gray identify the home location. The arrows show the movement of the mobile node in the static networks map. The movement of one mobile node in the sample map of cellular network consisting of 148 hexagonal cells is shown below in Figure 4.

Figure 7

Figure 9 Figure 4: Movement of a mobile node in the sample model network

Here it is assumed that the mobile station receives a call at every cell it visits. The following diagrams shows how a mobile station moves to the visiting cells. The system tracks the mobile station by the above mentioned relational location management concept. The diagrams 5 to 16 below reveal the fact that the number of paged cells can be reduced instead of

Figure 8

Figure 10

Figure 11

Figure 12

Figure 17: Informing cells in coverage area

Figure 13

Figure 14

Figure 18: Generalization of a network consisting of four composite cells

Figure 15

Figure 16

Figure 5-16: A Diagrammatic Discussion of TPRLM Process

E.

TPRLM Algorithm and Illustration

Now we present the TPRLM Algorithm after specifying the assumptions taken for its realization. We assume here that: a. The cells are of fixed shape. b. Some of the cells are special cells known as Informing Cells (IC) and the rest are ordinary MSC. c. The ICs can place request to an ordinary cell sharing its neighboring boundary for paging. d. No cell can be paged more than once in a single process. e. All ICs can only page cells of same Level in which they are or one level higher than it. f. The process ends when the mobile node is found and the location is updated. The ICs are positioned in balanced manner for any coverage area as shown in figure 17. Figure 18 shows the generalization of a network consisting of four composite cells. The cell ID in our scheme is represented using 8 bits. The left most 3 bits are used for the level numbers where as right most 5 bits are used for cell numbers in each level. Here we take the previous home location of any mobile node as H 0 .

The algorithm is written assuming 5 levels. The priority initialization in the algorithm is done based on priority calculation in table 1. The time interval of paging is determined and implemented according to the priorities calculated in table 2. The algorithm for each mobile node’s location management through TPRLM is given below. TPRLM_priorityinit () { If (priority ==1) Delay for 0.1 sec; Else If (priority ==2) Delay for 1 sec; Else If (priority ==3) Delay for 10 sec; Else If (priority ==4) Delay for 100 sec; Else Delay for 1000 sec; }

//when priority==5

TPRLM () { Step 1: Calculate the priorities of groups based on the inward call arrival rates and initiate the coverage area with ICs. Step 2: Call TPRLM_priorityinit(). n=1; Step 3: Search H 0 ; Step 4: If the mobile node found //based on cell ID Go to step 12.

Else go to step 5. Step 5: Request the ICs of the n level to search within itself; Step 6: If found, then go to step 12. Else go to step 7. Step 7: Search the cells in the common boundary of the same level; Step 8: If found, then go to step 12 Else go to step 9. Step 9: Search the cells in the common boundary of the ICs of (n+1)th level. Step 10: If found, go to step 12. Else go to step119. Step 11: If not found in the n level then the ICs will request to the ICs of the (n+1)th level and go to step 5; Step12: Update H 0 with the cell where the mobile node is Found and return; } In the following figures 19-24, the algorithmic process is illustrated diagrammatically for 3 level composite cell. The grey colored cells in these figures indicate paged cells.

Figure 23: Searching the 3rd level cells

F.

Figure 24: All cells are scanned

TPRLM: Discussion

Through invocation of time-based probabilistic concept and relational level-based coverage area management, location update delay, paging cost, paging delay, network bandwidth wastage etc. are reduced. The concept of the nth level composite cell enhances the generalization of network which improves the scope for user-defined flexibility. The number of informing cells (ICs) is also optimized and yet a fair balance in location management is maintained. Our attempt to achieve interactivity and enhanced performance is also very much desirable to cope with present requirement. The IC organization, movement management etc. support us to achieve our goals. VII. CONCLUSION

Figure 19: Searching the 1st level cells

Figure 20: Searching neighboring higher level cells

Figure 21: Searching the 2nd level cells Figure 22: Searching the neighboring higher level cells

Mobility management is an extremely important issue in today’s fast-upcoming and tremendously moving world. Location management, in this perspective, is one of the most vital affairs to be reflected upon. The fundamental concepts of location management and several techniques of locating any mobile node are explained in the earlier sections of this paper. Then various location management scheme proposals for wireless environment are surveyed to find out their advantages and development scopes. We have proposed the framework of the scheme TPRLM that looks for better interactivity, decreased location management cost and reduced location update and paging delay. It also attempts to reduce the network bandwidth wastage and improve the performance of location management for wireless cellular networks. We are unable to produce final simulation results in this paper. The work of simulation is presently going on. The algorithm may be augmented further to track other critical situations that are not covered in the present work. Analytical comparison of performance with other competent related proposals are another scope of future expansion of our work. REFERENCES [1] Demetres Kouvatos, Salam Adli Assi, Is-Haka Mkwawa, Vicente Casares Giner, Hermann Demeer and Amine Houyou, “ An Information-Theoretic Approach to Mobility Location Management: An Overview”, http://www.comp.brad.ac.uk/het-net/HET-NETs05/ReadCamera05/T10.pdf [2] Douglas E. Comer, “Internetworking with TCP/IP”, Volume 1, 1991 ? or http://www.leapforum.org/published/internetworkMobility/split

[3] Jun-Zhao Sun, Douglas Howie and Jakko Sauvola, “Mobility management techniques for the next generation wireless networks”, http://www.mediateam.oulu.fi/publications/pdf/85.pdf [4] Jingyuan Zhang, “Location Management in Cellular Networks”, in Handbook of wireless networks and mobile computing, pp. 27--49. John Wiley & Sons, Inc., 2002. [5] Fahd A. Batayneh, “Location Management in Wireless Data Networks”, http://www.cs.wustl.edu/~jain/cse574-06/ftp/wireless_location.pdf [6] A. Bar-Noy, I. Kessler and M. Sidi, “Mobile users: To update or not to update?” , Wireless Networks, 1 (1995) 175-185. [7] S. Ramanathan and M. Steenstrup, “A survey of routing techniques for mobile communication networks”, Mobile Networks and Applications, 1 (1996) 89-104. [8] H. Xie, S. Tabbane, and D. J. Goodman, “Dynamic Location Area Management and Performance Analysis”, Proceedings of 43rd IEEE Vehicular Technology Conference, Secaucus, NJ, May 1993, pp. 536-539. [9] C. Rose, “Minimizing the average cost of paging and registration: A timerbased methods”, Wireless Networks, 2 (1996) 109-116. [10] I. F. Akyildiz and J. S. M. Ho, “Dynamic mobile user location update for wireless PCS networks”, Wireless Networks, 1 (1995) 187-196. [11] Travis Keshab, “Location management in Wireless Cellular Networks” , http://www.cs.wustl.edu/~jain/cse574-06/ftp/cellular_location.pdf [12] J. S. M. Ho and I. F. Akyildiz, “Mobile user location update and paging under delay constraints”, Wireless Networks, 1 (1995) 413-425. [13] I-F. Tsai and R.-H. Jan, “The lookahead Strategy for Distance-Based Location Tracking in Wireless Cellular Networks”, ACM Mobile Computing and Communications Review, vol. 3, no. 4, 1999, pp. 27-38. [14] B. Liang and Z. J. Haas, “Predictive Distance-Based Mobility Management for PCS Networks”, INFOCOM, New York City, NY, USA, March 1999. [15] F. G. Nocetti, I. Stojmenovic, and J. Zhang, “Addressing and routing in hexagonal networks with applications for location update and connection rerouting in cellular networks”, from Internet. [16] S. K. Sen, A. Bhattacharya, and S. K. Das, “A selective location update strategy for PCS users”, Wireless Networks, 5 (1999) 313-326. [17] Y.-B. Lin, “Reducing Location Update Cost in a PCS Network”, IEEE/ACM Transactions on Networking vol.5, no.1, February 1997, pp.25-33. [18] J. G. Markoulidakis, G. L. Lyberopoulos, D. F. Tsirkas, and E. D. Sykas, “Evaluation of location area planning scenarios in future mobile telecommunication systems”, Wireless Networks, 1 (1995) 17-25. [19] W. Stallings, “Wireless Communications & Networks”, Prentice Hall of India Private Limited, New Delhi, 2003. [20] C. Perkins, “Mobile IP”, IEEE Communications Magazine, pp. 84-99, May 1997. [21] N. Venkataraman, “Inside Mobile IP”, Dr. Dobb’s Journal, Sep. 2005, pp. 24-30. [22] Sourav Saha and Sarmistha Neogy, “An Analytical Survey of Mobility Management Schemes for Wireless Networks”, unpublished. [23] B. Krishnamachari, R. Gau, S.B. Wicker and Z.J. Haas, “Optimal sequential paging in cellular wireless networks”, Wireless Networks. 10 (2004) 121–131. [24] H. Xie, S. Tabbane and D. J. Goodman, “Dynamic Location Area Management and Performance Analysis”, Proc. Of the 43rd IEEE Vehicular Technology Conference. Pp. 536-539, 1993. [25] Wenye Wang, Ian F. Akyildiz, Gordon L. Stuber, “Effective Paging Schemes with Delay Bounds as QoS Constraints in Wireless Systems”, Wireless Networks, 2001. [26] Dong-Jun Lee and Dong-Ho Cho, “On Optimum Timer Value of Area and Timer-based Location registration Scheme” , IEEE Communications Letters, Vol. 5, No. 4, April, 2001. [27] G.L.Lyberopoulos, J.G.Markoulidakis, D.V.Polymeros, D.F.Tsirkas, and E.D.Sykas, “Intelligent Paging Strategies for Third Generation Mobile Telecommunication Systems”, IEEE Transactions on Vehicular Technologies, vol. 44, no. 3, p. 543-553, August 1995. [28] Ian.F. Akyildiz, Joseph .S.M Ho, and Yi-Bing Lin, “Movement-based location update and selective paging for PCS network”, IEEE/ACM Transaction on Networking. 4 (4)(1996) pp. 629–638. [29] Vicent Casares-Giner and Pablo Garca-Escalle, “On the fractional movement-distance based scheme for PCS location management with selective paging”, First International Workshop in Wireless System and Mobility in Next Generation Internet. (2004) 202–218.

[30] A. Bar-Noy and I. Kessler, “Tracking Mobile Users in Wireless Communications Networks”, IEEE Transactions on Information Theory, vol. 39, no. 6, November 1993, pp. 1877-1886. [31] S. Tabbane, “An Alternative Strategy for Location Tracking”, IEEE Journal onSelected Areas in Communications, vol. 13, no. 5, June 1995, pp. 880-892. [32] J. Scourias and T. Knuz, “A Dynamic Individualized Location Management Algorithm”, Proc. 8th IEEE International Symposium on Personal, Indoor, and Mobile Radio Communications, Helsinki, Finland, September 1997, pp. 1004-1008. [33] A. Bar-Noy, I. Kessler and M. Naghshineh, “Topology-based tracking strategies for personal communication networks”, Mobile Networks and Applications, 1(1996) 49-56. [34] A. Bhattacharya and S. K. Das, “LeZi-Update: An Information-Theoretic Approach to Track Mobile Users in PCS Networks”, MOBICOM, Seattle, WA, USA, 1999, pp. 1-12. [35] H. Levy and Z. Naor, “Active tracking: Locating mobile users in personal communication service networks”, Wireless Networks, 5 (1999) 467-477. [36] Charles E. Perkins and David B. Johnson, “Mobility Support in IPv6”, Proc. Of MOBICOM, New York, USA, 1996. [37] A. Campbell, J. Gomez, C-Y. Wan, S. Kim, Z. Turanyi, and A. Valko, “Cellular IP”, Internet Draft, draft-ietf-mobileipcellularip-00, Dec. 1999. [38] Bernd Gloss and Christian Hauser, ”The IP Micro Mobility Approach”, Proc. of the EUNICE 2000 Open European Summer School on Innovative Internet Applications, Enschede, 2000. [39] C. Castellucia and L. Bellier, “Towards a unified hierarchical mobility management framework”, Internet Draft, draftcastellucia-uhmm-framework00,June 1999. [40] S. Das, A. McAuley, A. Dutta, A. Misra, K. Chakraborty and S. K. Das, “IDMP: An Intra-Domain Mobility Management Protocol for Next Generation Wireless Networks”, IEEE Wireless Magazine, October, 2002. [41] A. Misra, S. Das, A. McAuley, A. Dutta and S. K. Das, “IDMP: An intra-domain mobility management protocol using mobile agents”, Internet Draft, draft-misra-mobileip-idmp-00.txt,July. 2000. [42] K. Daniel Wang, Ilung-Yu Wei, Ashutosh Dutta, Kenneth Young, “Performance of IP Micro-Mobility Management Schemes Using Host Based Routing”, Proc. Of WPMC, 2001. [43] S. Das, A. Misra, P. Agarwal and S. K. Das, “TeleMIP: Telecommunication Enhanced Mobile IP Architecture for Fast Intra-Domain Mobility”, IEEE Personal Communications Systems Magazine, Vol. 7, No. 4, pp. 50-58, August 2000. [44] I. S. Misra, M. Chakraborty, D. Saha, A. Mukherjee, B. Deb, and B. Chatterjee, “A Three Level Hierarchical Mobility Management Architecture for IPv4 Networks”, Proceedings of International Conference IST Mobile and Wireless Summit 2004, pp. 358-362, June 2004, Lyon, France. [45] I. S. Misra, S. Dey and D. Saha, “TLMIPv6: A Next Generation Three Layer Mobile IPv6 Mobility Management Architecture”, Proc. Of World Wireless Congress, San Fransisco, May, 2005. [46] V. Casares-Giner and J. Mataix-Oltra, “On Movement-Based Mobility Tracking Strategy - An Enhanced Version”, IEEE Communications Letters, vol. 2, no. 2, February 1998, pp. 45-47. [47] H. Xie, S. Tabbane, and D. J. Goodman, “Dynamic Location Area Management and Performance Analysis”, Proc. 43rd IEEE Vehicular Technology Conference, Secaucus, NJ, May 1993, pp. 536-539. [48] Hongmin Chen and Nallasami Mani, “An Effective Scheme for Minimising Average Paging Costs Under Delay Bounds In PCS Networks”, http://atnac2003.atcrc.com/ORALS/Chen.pdf