Strategies for Cache Invalidation of Location Dependent Data in Mobile Environment

1
Strategies for Cache Invalidation of Location
Dependent Data in Mobile Environment
IRISS05 The 4th Annual Inter Research Institute
Student Seminar in Computer Science at I.I.T.
Kanpur, April 1-2, 2005

• By
• Ajey Kumar, Manoj Misra and A. K. Sarje

Department of Electronics and Computer
Engg. I.I.T. Roorkee
2
OVERVIEW
IRISS05

• Introduction
• Related Work
• Contribution
• Performance Evaluation
• Conclusion

3
IRISS05
INTRODUCTION
Mobile System Model
MH Mobile Host MSS Mobile Support Station Cell
Geographical Coverage Area under an MSS
4
IRISS05
INTRODUCTION contd.

• Characteristic of Mobile Elements
• Limited memory
• Limited computational power
• Small screen
• Limited battery Life
• Relatively Unreliable
• Variability in resources
• Frequent location updates

• Characteristic of Wireless Communication
• Frequent Disconnections
• Physical support for Broadcast
• Asymmetry
• Monetarily expensive
• Relatively unreliable
• High bandwidth Variability
• Low Bandwidth

5
IRISS05
INTRODUCTION contd.
Location- Dependent Information Services (LDIS)

• Information provided to users reflects its
  current location.
• Examples
• Advance Traveler Information System ( ATIS),
• GUIDE Project, etc.

6
IRISS05
INTRODUCTION contd.
Potential Applications

• Information Services
• Emergency Services
• Traffic Management, etc

7
IRISS05
INTRODUCTION contd.
LDIS Terminology

• Location Models
• Query Types
• Valid Scope

8
IRISS05
INTRODUCTION contd.
Location Model

• Geometric Model
• A location is specified as a 3-dimensional
  coordinate e.g. GPS.
• Symbolic Model
• The location space is divided into disjoint
  zones and each zone is identified with a unique
  name e.g. cellular infrastructure.

9
IRISS05
INTRODUCTION contd.
Query Types

• Mobile clients ,querying static objects
• Stationary clients, querying moving objects
• Mobile clients, querying mobile objects

10
IRISS05
INTRODUCTION contd.
Valid Scope The valid scope is defined as the
region within which the item value is
valid. Scope Distribution It is the set of
valid scopes for all items values of a data item.
11
IRISS05
INTRODUCTION contd.
What makes LDIS Challenging?

• Mobile Environment Constraints
• Spatial Property of Queries
• User Movement

12
IRISS05
INTRODUCTION contd.
Caching
13
IRISS05
INTRODUCTION contd.
Caching..

• Advantages of Data caching on mobile clients
• Improved access latency,
• Less wireless bandwidth requirements,
• Low energy/power consumption due to lower data
  transmission, and
• Improved data availability in case of
  disconnection

14
IRISS05
INTRODUCTION contd.
Cache Invalidation Methods

• Client-initiated Method
• The client monitors the states of the cached
  items and initiates the validity checking
  procedure.
• Server-initiated Method
• The server monitors the states of the cached
  items and informs the client to purge the
  obsolete data.

15
IRISS05
RELATED WORK
Location-Dependent Cache Invalidation

• (Based on Geometric Model)
• Polygon Endpoints (PE)
• Approximate Circle (AC)
• Caching Efficiency Based (CEB)

16
IRISS05
RELATED WORK contd.
Polygon Endpoints (PE)
At Server
At Client
17
IRISS05
RELATED WORK contd.
Approximate Circle (AC)
At Server
At Client
18
IRISS05
RELATED WORK contd.
Caching Efficiency Based (CEB)

• CEB is a generic method for balancing the
  overhead and the precision of valid scopes.
• The new performance criterion, caching efficiency
  of the data value with respect to a scope is
  defined as follows

19
IRISS05
RELATED WORK contd.
Caching Efficiency Based (CEB)

• For a data item value with valid scope of v,
  given a candidate valid scope set ,

choose the scope
that maximizes caching efficiency as the valid
scope to be attached to the data
20
IRISS05
RELATED WORK contd.
Given Data Items Valid Scope
c1
Candidates of Valid Scope Set
21
IRISS05
RELATED WORK contd.
Algorithm A1 Selection of the Best Valid Scope
for the CEB Method Input valid scope v p (e1,
,en ) of a data value Output the attached
valid scope v Procedure 1 v1 the
inscribed circle of p (e1, ,en ) 2 v v1
E max E(v1 ) 3 v2 p (e1, ,en ) 4 i
2 5 while n - i 1 do 6
//containing at least three endpoints for a
// polygon 7 if E(vi) gtE max then 8 v
vi E max E(vi ) 9 end if 10 if n - i
gt 1 then 11 vi1 the polygon that is
deleted one endpoint from vi
while being bounded by v an d has the maximal
area. 12 end if 13 i i 1 14 end
while 15 output v.
22
IRISS05
CONTRIBUTION
Generalized Caching Efficiency Based (CEB_G)
Algorithm A2 Selection of the Best Valid Scope
for the CEB_G Method Input valid scope v p
(e1, ,en ) of a data value Output the
attached valid scope v Procedure 1 v1
the inscribed circle of p (e1, ,en ) 2 v
v1 E max E(v1 ) 3 v2 p (e1, ,en )
4 i 2 5 while n - i 1 do 6
//containing at least three endpoints for a
// polygon 7 if E(vi) gtE max then 8 v
vi E max E(vi ) 9 end if 10 if n - i
gt 1 then 11 vi1 the polygon having
maximum area, consisting of ((n 1) i
2 ) endpoints of v and being bounded by v
12 end if 13 i i 1 14 end while
15 output v.
23
IRISS05
CONTRIBUTION contd.
Case Study
Best candidate for CEB_G
Best candidate for CEB
Original Polygon
24
IRISS05
CONTRIBUTION contd.
Stepwise Execution for CEB and CEB_G
25
IRISS05
CONTRIBUTION contd.
Stepwise Execution with best two in CEB
26
IRISS05
CONTRIBUTION contd.
Caching Efficiency with Future Access Based
(CEFAB)
Future Movement Path (FMP) for interval TQ, EMI
as
27
IRISS05
CONTRIBUTION contd.

CEFAB ..

v
v

v

v
v
Redefining FMP with respect to the valid scope v
for interval TQ, EMI, we have
28
IRISS05
CONTRIBUTION contd.
CEFAB ..
Goal is to select a valid scope that increases
the cache hit of the client, which means the sub
polygon which retains the total FMP.
New metric called Future Access (FA) for valid
scope vi for interval TQ, EMI, given by
where, vi sub region contained in v
vvalid scope of a data value
Length computes length of line segment between
two given end points.
29
IRISS05
CONTRIBUTION contd.
CEFAB ..
An integrated metric, Caching Efficiency with
Future Access (CEFA) for valid scope vi in
interval TQ, EMI, given by
30
IRISS05
CONTRIBUTION contd.
CEFAB ..
Algorithm A3 Selection of the Best Valid Scope
for the CEFAB Method Input valid scope v p
(e1, ,en ) of a data value, TQ and EMI Output
the attached valid scope v Procedure 1 if
TQ EMI then 2 vv 3
go to 19 4 end if 5
v1 the inscribed circle of p (e1, ,en ) 6
v v1 CEFA max E(v1 ) 7 v2 p (e1,
,en ) 8 i 2 9 while n - i 1 do
10 //containing at least three endpoints for
a // polygon 11 if
gtCEFA max then 12 v vi CEFA max 13
end if
31
IRISS05
CONTRIBUTION contd.
14 if n - i gt 1then 15vi1 the polygon
having maximum
,consisting of ((n 1) i 2 )
endpoints of v and being bounded by v 16 end
if 17 i i 1 18 end while 19 output
v.
32
IRISS05
SIMULATION MODEL
Features

• System execution model, client execution model
  and server execution model .
• we assume a ''wrappedaround'' model for the
  service area, represented by a rectangle of a
  fixed size.
• Scope distributions of the data items are
  generated based on voronoi diagrams and contains
  110 points randomly distributed in a square
  Euclidean space.
• The mobile client with fixed cache size is
  modeled with two independent processes query
  process and move process .
• Clients access pattern over different items
  follow a Zipf distribution.
• Clients wait for an exponentially distributed
  time period between successive query.
• The server is modeled by a single process that
  services the requests from clients on FCFS
  service principle.

33
IRISS05
SIMULATION MODEL contd.
Parameter for Server Execution Model
34
IRISS05
SIMULATION MODEL contd.
Parameter for Client Execution Model
35
IRISS05
SIMULATION MODEL contd.
Default Parameter Settings
36
IRISS05
PERFORMANCE EVALUATION

• The results are obtained when the system has
  reached the stable state, i.e., the client has
  issued at least 20,000 queries, so that the
  warm-up effect of the client cache is eliminated.
  
• The LRU cache replacement policy is employed for
  cache management.
• Cache hit ratio is employed as the primary
  performance metric. Specifically, the higher the
  cache hit ratio, the higher the local data
  availability, the less the uplink and downlink
  costs, and the less the battery consumption.

37
IRISS05
PERFORMANCE EVALUATION
38
IRISS05
CONCLUSION

• CEB_G maximizes the caching efficiency and show
  better performance as compared to existing CEB
  algorithm.
• Moreover, by varying CEB with more choices in
  each iteration, better results can be obtained.
• CEFAB gave the best performance
• As our future work we are extending our study for
  prefetching and Cache replacement policies for
  location dependent data.

39
IRISS05
Questions
?
Suggestions!!!
40
IRISS05
Thank You