Supporting Imprecision in Multidimensional Databases Using Granularities

1
Mobile Data Management
Instructor Sanjay Madria
Lesson Title - Location Management
2
Personal Communication System (PCS)

• A system where wired and wireless networks are
  integrated for establishing communication.

3
Personal Communication System (PCS)

• Wireless Components

4
Personal Communication System (PCS)

• Mobile cells
• The entire coverage area is a group of a number
  of cells. The size of cell depends upon the
  power of the base stations.

5
Personal Communication System (PCS)

• Problems with cellular structure

• How to maintain continuous communication between
  two parties in the presence of mobility?
• Solution Handoff

• How to maintain continuous communication between
  two parties in the presence of mobility?
• Solution Roaming

• How to locate of a mobile unit in the entire
  coverage area?
• Solution Location management

6
Personal Communication System (PCS)

• Handoff

A process, which allows users to remain in touch,
even while breaking the connection with one BS
and establishing connection with another BS.
7
Personal Communication System (PCS)

• Handoff
• To keep the conversation going, the Handoff
  procedure should be completed while the MS (the
  bus) is in the overlap region.

8
Personal Communication System (PCS)

• Roaming

• Roaming is a facility, which allows a subscriber
  to enjoy uninterrupted communication from
  anywhere in the entire coverage space.
• A mobile network coverage space may be managed
  by a number of different service providers. They
  must cooperate with each other to provide roaming
  facility.
• Roaming can be provided only if some
  administrative and technical constraints are met.

9
Personal Communication System (PCS)

• Roaming

Administrative constraints

• Billing.
• Subscription agreement.
• Call transfer charges.
• User profile and database sharing.
• Any other policy constraints.

10
Personal Communication System (PCS)

• Roaming

Technical constraints

• Bandwidth mismatch. For example, European 900MHz
  band may not be available in other parts of the
  world.
• Integration of a new service provider into the
  network. A roaming subscriber must be able to
  detect this new provider.
• Service providers must be able to communicate
  with each other. Needs some standard.
• Mobile station constraints.

11
Personal Communication System (PCS)

• Roaming

• Two basic operations in roaming management are
• Registration (Location update) The process of
  informing the presence or arrival of a MU to a
  cell.
• Location tracking the process of locating the
  desired MU.

12
Personal Communication System (PCS)

• Roaming

• Registration (Location update) There are six
  different types of registration.
• Power-down registration. Done by the MU when it
  intends to switch itself off.
• Power-up registration. Opposite to power-down
  registration. When an MU is switched on, it
  registers.
• Deregistration. A MU decides to acquire control
  channel service on a different type of network
  (public, private, or residential).

13
Personal Communication System (PCS)

• Roaming

• Registration (Location update) There are six
  different types of registration.
• New system/Location area registration when the
  location area of the MU changes, it sends a
  registration message.
• Periodic registration A MU may be instructed to
  periodically register with the network.
• Forced registration A network may, under certain
  circumstances, force all MUs to register.

14
Personal Communication System (PCS)

• Registration

Two-Tier Scheme

• HLR Home Location Register
• A HLR stores user profile and the geographical
  location of each moving object at a pre-specified
  location
• VLR Visitor Location Register
• A VLR stores user profile and the current
  location who is a visitor to a different cell
  than its home cell.

15
Personal Communication System (PCS)

• Registration

Two-Tier Scheme steps. MU1 moves to cell 2.
16
Personal Communication System (PCS)

• Registration

• Steps
• MU1 moves to cell 2. The MSC of cell 2 launches
  a registration query to its VLR 2.
• VLR2 sends a registration message containing MUs
  identity (MIN), which can be translated to HLR
  address.
• After registration, HLR sends an acknowledgment
  back to VLR2.
• HLR sends a deregistration message to VLR1 (of
  cell 1) to delete the record of MU1 (obsolete).
  VLR1 acknowledges the cancellation.

17
Personal Communication System (PCS)

• Location tracking (MU2 wants to comm with MU1)

• Steps
• VLR of cell 2 is searched for MU1s profile.
• If it is not found, then HLR is searched.
• Once the location of MU1 is found, then the
  information is sent to the base station of cell
  1.
• Cell 1 establishes the communication.

18
Locating Objects in Mobile Computing
Location Management Lookups and updates
19
Introduction

• Our target is objects capable of changing their
  location
• We are interested in objects with identity
• We store user locations in multiple databases
  (DBs)
• Main questions
• How do we update data when user moves?
• How do we locate user in DBs when it is required?

20
Locating Moving Objects

• Example of moving objects
• mobile devices (cars, cellular phones, palmtops,
  etc)
• mobile users (locate users independently of the
  device they are currently using)
• mobile software (e.g., mobile agents)
• How to find their location - Two extremes
• Store their current location everywhere
• Search locally
• Cost of updates
• Search everywhere
• No information is stored anywhere search is
  expensive
• No cost of updates
• Searching verses Update Cost

21

• Availability either at all sites or at
  selective sites (frequently visited sites)
• Currency Stored location is always updated (it
  may not make sense if user moves very frequently)
• Precision Exact location verses set of possible
  locations

22
Locating Moving Objects

• What (precision), where (availability) and when
  (currency) to store

at all sites
Availability
At selective sites (e.g., at frequent callers)
the whole network
nowhere
Exact location
Set of locations
Currency
Precision
Never update
Always update (at each movement)
23
Overview

• Database schemas
• Two-tier
• Hierarchical
• Replication
• Working set replication
• Replication in hierarchical schema
• Forwarding pointers
• Two-tier schema
• Hierarchical schema

24
Architectures of Location DBs

• Two-tier Schemes (similar to cellular phones)
• Home Location Register (HLR) store the location
  of each moving object at a pre-specified location
  for the object
• Visitor Location Register (VLR) also store the
  location of each moving object at a register at
  the current region
• Hierarchical Schemes
• Maintain multiple registries

25
Two Tier Scheme

• HLR Home Location Register is associated with
  each mobile user, maintains current location of
  the user as part of the users profile
• HLR is located at a network location
  pre-specified for each user
• To locate X, Xs HLR is queried
• When X is moved to new zone, Xs HLR is contacted
  and updated
• Disadvantages Global move is expensive

26
Two-tier Location DBs

• Search
• Check the VLR at your current location
• If object not in, contact the objects HLR
• Update
• Update the old (delete) and new VLR (insert)
• Update the HLR

27
Two-tier Schema- Enhancement

• Users X profile is permanently stored in Home
  Location Register (HLR)
• Each site maintains Visitor Location Register
  (VLR), stores info about users not at their home
  location
• During lookup at VLRi
• VLRi is queried for X location
• HLR of X is queried upon failure
• During move from i to k
• Xs HLR is updated
• Xs profile is deleted from VLRi
• Xs profile is added to VLRk

VLRHLR-X
VLRiHLRi
VLRk-XHLRk
28
Two-tier Schema Cont.

• Does not support locality
• search in nearby locations impossible, always
  need to reg. with HA
• possible distant HLR is always contacted upon
  move
• Home Location register is permanent
• resettlement is not supported
• Does not scale well Home location is always
  contacted
• Relatively simple
• max 2 operations for lookup
• 3 operations for update

VLRHLR-X
VLRiHLRi
VLRk-XHLRk
29
Standards

• EIA/TIA (Electronic Indus Asso.) and GSM(Global
  System for mobile Comm.) Use HLR and VLR
• Mobile-IP - Two IP addresses Home address and
  Care-of Address (Current point of attachment)
• Care-of-address is either the address of FA or IP
  address acquired by the node in the current
  network
• Mobile node registers its care of address with
  its home address

30
Hierarchical Schema

• A hierarchy of location databases is maintained
• Internal node maintains information about user
  registered in the set of zones in its subtree
• Leaf node contains actual location of objects in
  its coverage
• Intermediate node contains location information
  for all objects covered by its children in a form
  of
• Pointers to lower level DBs
• or
• Actual location of each object

1
2
3
7
6
5
4
18
19
20
15
16
17
12
13
14
8
9
10
31
Hierarchical Scheme

• LCA(I,j) least common ancestor of nodes I and j
• Parameter that affect the performance of the most
  location management scheme
• Relative frequency of the move
• Call operations of each user
• Call to mobility ratio (CMRi) Ci/Ui
• Where Ci is the expected number of calls to user
  Pi over a period T and Ui the number of moves
  made by Pi over a period T
• LCMRij (local call to mobility ratio involves
  origin of calls)
• Cij/Ui
• For hierarchical scheme, LCMRij ? LCMRik where
  k is a child of j the call to mobility ratio for
  a user Pi and an internal node j is the ratio of
  the number of calls to Pi originated from any
  zone at Js subtree to the number of moves made
  by Pi

32
Lookup in Hierarchical Schema

• Whenever lookup for object X at i is initiated
  at j
• The tree is traversed from j upwards to LCA(i,j)
• Then
• Pointers are traversed downwards from LCA(i,j) to
  i
• Location of X is found at i
• (pointer case)
• OR
• Location of X is found in LCA(i,j)
• (actual location case)
• LCA(i,j) Least Common Ancestor of i and j
• Example (j,i) (8,12), (19,16)

1
LCA(19,16)
LCA(8,12)
2
3
7
6
5
4
18
19
20
15
16
17
12
13
14
8
9
10
33
Moving in Hierarchical Schema

• Whenever object X moves from j to i
• Pointers on the path j,,LCA(j,i),i are altered
• OR
• All databases on paths root,,j and root,,i are
  updated with new location of X

14
1
14
2
3
14
7
6
5
4
18
19
20
15
16
17
12
13
14
8
9
10
X
Y
34
updates

• Pointer case - When user x moves from 15 to 18,
  the entries at 18,7, 6, 3, 15 are updated x is
  deleted from the database at 15 and 6, at 3 it is
  updated and are added at 18 and 7
• Actual location case entries are updated at 0,
  7,6,3,18,15

35
Evaluation of Hierarchical Schema

• Mobile object is not bound to HLR
• Advantage of locality moves and lookups is taken
• Increased number of operations
• DB operations
• Communication messages
• Increased load and storage requirements for DBs
• Intermediate DBs store location information for
  all objects covered by its children
• Root DB stores location information for ALL
  objects

36
Partitions

• Avoid maintaining locations at all levels, reduce
  search cost
• Partitions grouping zones for each user among
  which it moves frequently
• Partition exploits locality
• Information whether a user is currently in the
  partition is maintained at the LCA of all nodes
  in partition called representative of the
  partition
• Representative does not know exact location
• Reduces overall search cost, increases update
  cost when a user crosses partition, both
  representatives must be informed

37
Locating Moving Objects
Partitions
P3
P4
P5
P1
P2
User x
User x
38
Overview

• Database schemas
• Two-tier
• Hierarchical
• Replication
• Working set replication
• Replication in hierarchical schema
• Forwarding pointers
• Two-tier schema
• Hierarchical schema
• Other topics
• Relation to KDE3 project
• Evaluation

39
Locating Moving Objects

• Caching
• cache the callees location at the caller side
• (large Call to Mobility Ratio)
• Replication
• replicate the location of a moving object at its
  frequent callers (large CMR)
• Forwarding Pointers
• do not update the VLR and the HLR, leave a
  forwarding pointer from the old to the new VLR
  (small CMR)
• When and how forwarding pointers are purged?

40
Caching Two Tier Scheme

• Current location of the callee may be reused by
  subsequent calls originated from same region
• Every time a user x is called, its location is
  cached at the VLR in the callers zone so it can
  be reused
• Caching is useful for those users who frequently
  receives calls relative to the rate at which they
  relocate
• To locate a user, the cache at the VLR of the
  callers zone is queries first, if found then
  query is launched to the indicated zone without
  contacting the users HLR else HLR is queried

41
Caching Hierarchical scheme

• When a call is made from zone i to user x located
  at zone j, the search traverse from i to LCA
  (i,j) and then down to j, Ack is returned back to
  i from j
• Forward and Reverse bypass pointers
• Forward bypass is an entry at an ancestor of i,
  say s, that points to ancestor of j, say, t
• Reverse is from t to s
• During the next call from zone i to user x,
  search will travel until s then follow t via LCA
  (i,j) or via a shorter path
• In case of simple caching, s and t can be at the
  leaf while in level caching, s and t belongs to
  any level and possibly to a different one.
• Placing a bypass at higher level node s makes
  this entry available to all calls made in ss
  subtree, but will travel long path to reach s
• Placing high node t will increase lookup cost,
  cache entry remains valid as long as the user
  moves inside ts subtree

42
Cache Invalidation

• Eager caching Every time a user is moved new
  location, all cache entries for this users
  location are updated
• Location of the cache entries must be centrally
  known, failure of central location can cause
  problem
• Lazy caching a move operation does not mean
  updating cache
• When look up, either the user is still in the
  indicated location or it has moved out (cache
  miss)
• Cache miss- HLR is contacted and then cache is
  updated
• Cache update only on cache miss
• Overhead cache location must be visited first
• Saving of lazy caching over eager caching if hit
  ratio threshold for a user in a zone must exceed
  the (cost of lookup when there is a hit) / (the
  cost of lookup in non-caching scheme). Depends on
  querying HLR and VLR

43
Replication

• Location (profile) of selected users is
  replicated at selected sites
• Enhances lookup response time
• Reduces network load during lookup
• Creates overhead during updates
• Replication is judicious if following holds
• number of lookups for object i from location
  j during T
• number of updates for i during T
• - savings per lookup - cost per
  update
• Additional parameters service capacity of DBs,
  etc.
• Replication is possible at both caller and
  receiver locations

(1)
44
Working Set Replication

• Applicable to two-tier schema
• Replicas are kept at frequent callers of X
  working set of X
• Equation holds for every member j of the set
• Every time a call to X is made
• From a member of a set
• no updates required
• From nonmember k of a set
• if (1) holds for k, k is added to the set
• Every time X moves
• (1) is evaluated for each member of X working set
• If (1) does not hold for member k, it is removed
  from the set

X
5
1
(1)
4
2
6
7
8
3
9
45
Replication in Hierarchical Schema

• Takes advantage of locality in movement
• Local Call to Mobility Ratio is used to determine
  feasibility
• and are thresholds for determining
  replication of nodes
• If , j is assigned a
  replica
• If , j is not used for
  replication
• If , database
  depends on topology of network

1

3
2



7
6
5
4
18
19
20
15
16
17
12
13
14
8
9
10

46
Overview

• Database schemas
• Two-tier
• Hierarchical
• Replication
• Working set replication
• Replication in hierarchical schema
• Forwarding pointers
• Two-tier schema
• Hierarchical schema

47
Forwarding Pointers (two-tier schema)

• Pointers could be used to reduce communication
  overhead and query load at the HLR
• When X moves from i to j a pointer from VLR at i
  to VLR at j is added
• During lookup if no information on X is found at
  current VLR, HLR of X is queried and pointers are
  followed
• Chain of pointers is managed not to exceed length
  K
• Useful for users receiving calls infrequently and
  reallocate often
• It has been showed that cost savings for Klt5,
  are 20-60

VLRn-XHLRn
48
Forwarding Pointers (hierarchical schema)

• Updating and lookup in hierarchical schema lack
  efficiency when LCA(i,j) is at higher levels
• Consider schema with entries as pointers to lower
  level DBs
• In simple forwarding a forwarding pointer
  connects two leaf nodes
• Allows cheaper updates
• In level forwarding a forwarding pointer connects
  two intermediate nodes
• Allows cheaper lookups

Level forwarding
Simple forwarding