Location management in Mobile Networks

ByC.Manoj Kumar

T. Aswin Kumar B. Anuradha

The Problem To track the mobile user Maintaining the binding between the logical

identifier and physical location of the user The location of the terminal cannot be deduced

from its endpoint address, like in wired networks Additional addressing schemes protocols needed

to locate the mobile terminal Being mobile, user creates uncertainty of the

exact location of the mobile terminal which is tried to overcome by the network

Location management Involves two basic operations

Paging Search by system to track the mobile MSC broadcasts message on the FCC Target replies in the reverse channel

Update Upper bound on the location uncertainty Mobile sends update message on the

reverse channel

Issues Cost of Communicating with mobile user is augmented by “Cost of Searching for the Current location of the user”.

Paging cost Number of calls arrived & Number of cells paged

Update cost Number of times mobile updates

Trade-off : More the paging, less the update Requirement for an optimal algorithm

Call routed with allowable time constraint Less information exchange

Update Schemas The vicinity of the last update information gives

the most probable location of the mobile terminal When to update? Static

Partition of cells in LAs Non-overlapping grouping of cells Mobile updates when it crosses an LA

boundary Drawback: traffic generated only on the

boundary cells, reducing BW availability for other calls

Update Schemas Selection of designated reporting cells

Mobile must update in some designated cells Optimal set of reporting cells is NP-Complete

Dynamic – based only on user’s activity Distance based

Updates when Euclidean distance crosses a threshold D

Distance can be specified in terms of cells covered Movement based

Updates when number of cell boundaries crossed reaches a threshold M

Time based Mobile sends periodic updates

a b c

hg

f e

d

Dynamic Update Schemes

Location Databases Distributed DBs used to store the

location of mobile users Types of Architectures

Two-tier Hierarchical Regional Directories

Two-Tier Architecture Used in GSM A home DB called “Home Location Register (HLR)” is

associated with each mobile user “Visitor Location Register (VLR)” is maintained at

each zone HLR

Located at a pre-specified zone for each user Maintains the current location of the user as part

of the “User Profile” To locate user ‘x’, x’s HLR is identified and queried When ‘x’ moves, contacts HLR and updates it to

“New Current Location”

VLR stores profiles of users not at home location

and currently located inside it’s area. When call is placed from zone ‘i’ to user ‘x’

Query for ‘x’ in i’ s VLR If ( not found )

contact x’s HLR When ‘x’ moves from zone ‘i’ to zone ‘j’

Delete entry ‘x’ from i’s VLR Add new entry ‘x’ to j’s VLR Update x’s HLR

Two-Tier Architecture (contd..)

Location Management in GSM The Mobility Management layer (MM) is built on top of

the RR ( Radio Resources Management ) layer. Handles the functions that arise from the mobility of the

subscriber, as well as the authentication and security aspects.

A powered-on mobile is informed of an incoming call by a paging message sent over the PAGCH channel of a cell.

Design Choices Page each cell in the network for each call

Waste of Bandwidth Page exactly one cell but requires the mobile to send

updates each time it changes cell Results in a large number of updates

Design Choice in GSM compromise solution

used in GSM is to group cells into “Location areas”

Updating messages are required when moving between location areas, and mobile stations are paged in the cells of their current location area.

Mobility Management (contd..) Location updating

When mobile powered ‘ON’ Performs update indicating it’s IMSI( International Mobile Subscription Id ) The above procedure called “ IMSI Attach Procedure “

When moves to new Location Area or a Different PLMN Update message sent to new MSC/VLR If MS authorised in the new MSC/VLR

thensubscriber’s HLR updates the cuurent location sends a message to the old MSC/VLR to cancel it’s VLR

entry

Periodic Location Updating If after the updating time period, the mobile station has not registered, it is then deregistered

When a mobile station is powered off it performs an “IMSI detach procedure” in order to tell the network that it is no longer connected.

       

MSC 1

To otherMSCs

HLR V

LR

MSC 2

HLR V

LR

MSC 3

HLR

VLR

USER INFO

CL

A MSC2

USER INFO

A

USER INFO

A

VLR

VLR

HLR

Fixed Network

Drawbacks of HLR/VLR Assignment of HLR to a mobile is

permanent users permanently shifted to different

region still contact the same HLR Scalability

Not scalable to highly distributed systems

Locality of mobility is not taken advantage of

Hierarchical Architecture Extend the two-tier scheme by

maintaining a hierarchy of location DB

Location DB at higher level contains location information for users located at levels below it

Types Static Adaptive

Static LMT Location mgt involves

Updates Searches Search-updates

A LMS is a combination of all the above strategies

Cost of LMS includes number, size and the distance message needs to travel

An efficient LMS should attempt to minimize the above cost

System Model Mobile networks comprise

. Static backbone network

. A wireless network Two distinct set of entities exist.

. Mobile hosts

. Fixed hosts Notion of Mobile Support Station(MSS) is there

which stands as a gate way between wireless and static network.

A Mobile host can communicate with only one MSS at a time.

Logical Network Architecture.Each LMS assumes following logical architecture for the system-

• MSS at the leaves• Location Server at the intermediate nodes• Every Mobile host(MH) will have an unique identifier called Home Address(HM). But the Physical addresses change

The aim of the location management strategy reduces to efficiently maintaining HM-PA mappings

Data Structures.

Every node (MSS/LS) maintains the information of mobile hosts as three tuples-

(h-id, fp-dest, fp-time). h- id : Mobile host identifier (Home Address). fp-dest : Forwarding pointer indicating dest.. fp-time : Last time where forwarding pointer

update took place. At the MSS, the fp-time value of a host residing in its

cell is NULL. Initial conditions: Every LS should have correct

information regarding the MHS residing in its subtree.(Root contains information about all MHS).

Update Protocols Strategies for updating the location information at MSS

and LS when MH moves LAZY-UPDATE(LU)

- Simple, updates take place only at the source and dest. - S-{h, D, t} , D-{h, dest, NULL)- Cost of update is ZERO.

FULL UPDATE(FU)- Location updates take place in all the LS’s located on the path from Source and destination cells to the root.- Cost is 2(h-1) where h is the height of the tree.

LIMITED UPDATE(LMU)- Update take place in the LS’s whose height is less than m(m<H).- Cost is 2(m-1).

Full Update algorithm (FU) Source cell

At the MSS: for host h set fp_dest = dest and fp_time =t

The MSS of src sends update messages to all the Location Servers upon which an entry (h,dest,t1) is added (t1 – local time of the location server)

Destination cell At the MSS: An entry (h,dest,NULL) is added for

host h MSS of the dest sends update messages to all the

location servers upon which an entry (h,dest,t1) is added (t1 – local time of the location server)

Search ProtocolInitially, the search_cell is the source cell CSTEP 1 :

If the MSS of the search_cell has an entry for h1If fp_dest= search_cell, search completed.else search_cell = fp_dest. Repeat step1

Else Forward the query to the next higher level LS.

STEP 2 :If the location server has an entry (h1, fp-dest, fp-time) for h1

search_cell = fp_dest. Go to step 1Else If the location server is the root,

Root broadcasts to find out the location of h1.Else Forward the query to the next higher level LS along the root and go to STEP1.

Search Update Protocols Make sense to perform updates at certain

location servers after a successful search Example

There are three search-update protocols.

Search-update protocols cont. NO-UPDATE(NU)

. As name indicates, no update-action will be taken after successful search.

. But, the fp-time field value updated in the search path.

JUMP UPDATE(JU). Location update takes place only at the callers MSS.. Update cost is 1.

PATH COMPRESSION UPDATE(PCU). Location update takes place at all the nodes in the

search path.. Cost is the length of the search path.

Purging of the forwarding pointers. There is a need to purge the forwarding pointers for two

reasons.- Save storage- Avoid storing stale information

So, the forwarding pointers are purged periodically.THE PROBLEM ! As LU and LMU do not update forward pointers at higher

levels, they will be purged as they become stale. But the problem is that root LS need to do frequent broadcasts.

The solution is that, along with purging, forwarding pointers at the LS on the path to the root from the current MSS must be updated periodically.

Simulations Location Management Strategy can be

simulated on two type of environments. Environment where the moves of mobile and also

the calls for it are arbitrary. Environment where short moves and a set of

callers for a mobile is assumed.(A realistic assumption).

PARAMETERS FOR THE TWO TYPES. Time b/n the moves of a host and calls for it are

assumed to be an exponential distribution with mean C and M respectively.

Height of the tree is chosen based on some probability distribution function

Observations Type I

FU - LMU : high cost of updates upon each move LU – NU : high search cost LU – JU : update on successful search does not reduce

search cost Type II

LU – JU outperforms all the other strategies as employing JU at the caller’s end reduces the search cost

Adaptive LMT System designers do not always have prior knowledge of

the call-mobility patterns for a particular host. In these cases, there is a need for a mechanism which can

dynamically adapt most suitable LMS from time to time.

DATA STRUCTURES

M(h) : {m1,m2,m3…mn }, mi ={ti,src,dest}—sequence of moves

Cu(h) : {cu1, cu2... cun }, cuj =cost of update upon move mj

S(h) : {s1, s2,.. sn }, si ={tsi,h1) there is call from h1 at tsi.

Cs(h) : {cs1, cs2,…. csn} – sequence of costs.

HOW THESE DATA STRUCTURES ARE OBTAINED?

Finding mobility and call frequencies FINDING MOBILITY.

A system parameter MTMI(Maximum Threshold Move Interval) is defined. Let Delta be the average time interval b/n successive moves.

Then if Delta<MTMI, the host is fast moving otherwise slow moving.

FINDING CALL-FREQUENCY.Again here, MTCI(Maximum Threshold Call Interval) is defined. Let Delta1 be average time interval b/n successive calls for the host.

Then if Delta1<MTCI, then the host is frequent caller otherwise in-frequent caller.

Adaptive algorithm If(host makes a lot of long moves)

Employ LU-PCelse if(Frequent caller and low mobility)

Employ LU-JUelse if(Frequent caller and high mobility)

Employ LU-PCelse if( Not frequently called and high mobility)

Employ LU-JU.else Employ LU-PC

Pros and Cons of Hierarchical Approach

Conclusions The Location DBs must support very high

update rates. Various enhancements to these approaches

include caching, replication etc.., Some approaches also try to predict the

probability of a user to be in a given cell based on it’s mobility pattern.

Support of advanced queries involving the location of moving objects is an ongoing research topic.

References Static and Adaptive Location Management in

Mobile Wireless Networks, P. Krishna, Nitin H. Vaidya and Dhiraj K. Pradhan,

LeZi-Update: An Information-Theoretic Approach to Track Mobile Users in PCS Networks, MOBICOM'99, Seattle

Locating Objects in Mobile Computing, Evaggelia Pitoura and George Samaras,

Efficient and Flexible Location Management Techniques for Wireless Communication Systems, Jan Jannik, Derek Lam, N Shivakumar, Jennifer Widom , Donald C. Cox