Lecture 4 mobile database system

MObile DBMS

1. Mobile
Databases

3. PowerpointTemplates
Why MobileDatabases?
Page 3
Numberofsmartphonesin usearoundthe world
passed 6.387billion in 2021.
Nextbilliondevicescouldbereachedwithinless
thanthreeyears.
Morebusinessesmovetowardemployees
mobility.
Powerfullightweightcomputingdevicesandlow
costmobileconnectivitypavedthewayfordata-
drivenapplications.

4. PowerpointTemplates
Why MobileDatabases?
Page 4
Mobile data-driven applications enable us to access
any data from anywhere, anytime.
Examples:
Salespersons can update sales records on the move.
Reporters can update news database anytime.
Doctors canretrieve patient’s
medical history from anywhere.
Mobile DBMSs are needed to support these applications
data processing capabilities.

5. PowerpointTemplates
MobileDatabase:
Page 5
Amobiledatabaseisadatabasethatcanbeconnected
tobyamobilecomputingdeviceoverawireless
mobilenetwork.
Mobile databases:
Physicallyseparate fromthecentral
database server.
Resided on mobile devices.
Capable of communicating with a central database
server or other mobile clients from remote sites.
Handle local queries without connectivity.

6. PowerpointTemplates
Client-Server MobileDatabases:
Page 6
`Central
Databas
e
Centra
l
DBMS
Mobile
DBMS
Mobile
DB
Mobile
DB
Mobile
DBMS
Centra
l
Server
Lapto
p
Smartphon
e

7. PowerpointTemplates
Client-servermodelisthetraditionalmodelof
information systems.
It is the dominant model for existing mobile databases.
Theservercanbecomeasinglepointoffailureand
performancebottleneck.
Evenstoringdataonaclusterofmachinestobackup
centraldatabasemightcauseperformancebottleneck
anddatainconsistency.
Client-Server MobileDatabases:
Page 7

8. PowerpointTemplates
Peer-to-Peer MobileDatabases:
Page 8
`
Mobile
DB
Mobile
DBMS
Mobile
DBMS
Mobile
DB
Mobile
DBMS
Mobile
DB
Lapto
p
Smartphon
e
Tablet

9. PowerpointTemplates
Peer-to-Peer MobileDatabases:
Page 9
InP2Pmobiledatabases,thedatabasemaintenance
activities are distributed among clients.
Everyprocessplayspartoftheroleoftheserver,
besidesitsclientrole.
Aclientthatwantstoaccessapieceofdata,sendsa
requesttootherpeerclientsandtheyforwardthe
requestuntilthedataisfound.
Themajorprobleminthismodelisensuringthe
availabilityofdata.

10. PowerpointTemplates
Characteristics of MobileEnvironments:
Page 10
Whynotuseanappropriateexistingmodelof
databasesinamobileenvironment?
Characteristicsofmobileenvironments:
Restricted bandwidth of wirelessnetworks.
Limited powersupply.
Limitedresources.
Mobility.
Disconnections.

11. PowerpointTemplates
CurrentApproach:
Page 11
Currently most mobile application developers use “flat
files” to store application data.
A “flatfile”is a filecontaining
records thathaveno structured
interrelationship.
Advantages:
Smaller and easier to manage.
Disadvantages:
Applicationsneedto knowthe organizationof
the
records within the file.
Developers haveto implement the required
database functionalities.

13. PowerpointTemplates
Requirements of MobileDBMSs:
Page 13
MobileDBMSsshouldsatisfythefollowing
requirements:
Small memoryfootprint.
Flash-optimized storagesystem.
Datasynchronization.
Security.
Low powerconsumption.
Self-management.
Embeddable inapplications.

14. PowerpointTemplates
Small MemoryFootprint:
Page 14
theoverall
Mobile DBMSs should be customizable to include only
the required database functionalities.
Memoryfootprintisamountofmain
memorythatanapplicationuseswhile
running.
Mobiledeviceshavelimitedmemory,so
themobiledatabaseapplicationshould
haveasmallfootprint.
Thesizeof mobiledatabase
affects
application footprint .

15. PowerpointTemplates
Flash-Optimized StorageSystem:
Page 15
Flashmemoriesaredominantstoragedevicesfor
portable devices.
They have feature suchas:
Smallsize.
Better shockresistance.
Low powerconsumption.
Fast accesstime.
No mechanical seek and rotationallatency.
MobileDBMSsneedtobeoptimizedtoexploitthe
advantages of the new storagedevices.

16. PowerpointTemplates
DataSynchronization:
Page 16
Portabledevicescannotstay connected all
the time.
Users can accessandmanipulate data
on their devices.
They are also unable to store a large amount of data
due to lack of storage capacity.
MobileDBMSs shouldhavethesynchronize
functionalitytointegratedifferentversionsofdatainto
aconsistentversion.

17. PowerpointTemplates
Security:
Page 17
Securityisveryimportantfordata-centricmobile
applications.
Itismoreimportantwhentheapplicationworkswith
criticaldatathatitsdisclosureresultsinpotentialloss
ordamage.
Data that are transmitted over a wireless network are
more prone to security issues.
Mobile DBMSs should implement a
complete end-to-
end security to ensures the secure transfer of data.

18. PowerpointTemplates
Low PowerConsumption:
Page 18
Portabledeviceshavelimitedpowersupplies.
Batterylifeofmobilephonesisexpectedtoincrease
only20%overthenext10years.
Processor,display andnetwork
connectivityarethe
main power consumers in a mobile device.
Mobile DBMSs need to be optimized for efficient power
consumption.

19. PowerpointTemplates
Self-Management:
Page 19
Intraditionaldatabases,thedatabase
administrator(DBA)isresponsiblefor
databases maintenance.
In mobile DBMSs there can be no DBA to manage
the database.
Mobile DBMSs need to support self-
management and automatically perform the DBA
tasks.
SomemobileDBMSsallowremotemanagementthat
enablesaDBAtomanagethemobiledatabasesfroma
remotelocation.

20. PowerpointTemplates
Embeddable in applications:
Page 20
Administratorsdoesnothavedirectaccesstomobile
devices.
MobileDBMSsshouldbeanintegralpartofthe
applicationthatcanbedeliveredasapartofthe
applications.
The database must be embeddable as a DLL file in the
applications.
Itmustbe alsopossible to deploy
thedatabase as a
stand-alone DBMS with support of multiple transaction.

21. PowerpointTemplates
Existing MobileDatabases:
Page 21
Mobiledatabases:
Sybase SQLAnywhere
OracleLite
Microsoft SQL ServerCompact
SQLite
IBM DB2 Everyplace(DB2e)
Embeddeddatabase:
TinyDB
PicoDBMS

22. PowerpointTemplates
Sybase SQLAnywhere:
Page 22
Initially created by Watcom as Watcom SQL.
SQL Anywhere was launched in in 1995.
It dominatesthemobile-database field,
with about 68% of the mobile database market.
Database files are independent of the operating
system
and transferable between supported platforms.
Strong encryption is supported for both database files
and client-server communication.

23. PowerpointTemplates
OracleLite:
Page 23
OmniscienceObjectTechnology,Inc.wasacquiredby
OracleCorporationinNovember1996.
Theirproduct(OmniscienceORDBMS)becamethefirst
versionofOracleLite.
OracleLiterunsinunder1MBofmemory,andcanbe
installedin3MBofharddiskspace.
PersonalOracleLite(POL)isalightweight,single-user
relationaldatabasethatrunsondesktops,laptops,
downtothesmallesthandhelpdevices.

24. PowerpointTemplates
Microsoft SQL ServerCompact:
Page 24
FormerlyknownasSQLServerMobile Edition.
SQL Server Compact is free to download and
redistribute.
It is optimized for an architecture where all applications
share the same memory pool.
SQL CE runs in-process with the application which is
hosting it.
It has a memory footprint of approximately 5 MB and
disk footprint of less than 2 MB .

25. PowerpointTemplates
SQLite:
Page 25
SQLite is an open source mobile database engine.
Itis a server-lessdatabase engine
that needs zero-configuration.
SQLite is a popular choice as mobile database for local
storage in mobile applications.
SQLite engine has no standalone processes with which
the application program communicates.
SQLite implements most of the SQL-92 standard.

26. PowerpointTemplates
IBM DB2Everyplace:
Page 26
DB2e has been discontinued and April 2013
is the end of support date.
KB)in
IthadthebiggestmarketshareafterSQL
Anywhere.
Ithadthesmallestmemory footprint
(350
compare to other commercial mobile databases.
data
IBM has replaced DB2e with IBM solidDB family.
SolidDBis a in-memory MDBMS with
robust catching features.

27. PowerpointTemplates
EmbeddedDatabases:
Page 27
Embeddeddatabasesystemsaretightlyintegrated
with an application that requires access to stored data.
They are hiddenfrom the
application’s end-userand
requires little or no ongoing maintenance.
Embedded databases need less resources in compare
with mobile databases.
They are optimizedforspecificdevices
such as smartcards and sensors.
They support limited and specified
functionalities of
the standard SQL.

28. PowerpointTemplates
PicoDBMS:
Page 28
PicoDBMSonlysupportssufficientfunctionalitiesfor
smartcardapplications.
Smartcardapplicationsareusedfordatamanagement
such as insert, delete, update andsearch.
PicoDBMS supports a part ofSQL:
INSERT, UPDATE, DELETE,SELECT
CREATE/DROP TABLE/VIEW
GRANT/REVOKE
Footprint size of PicoDBMS is about30KBytes.

29. PowerpointTemplates
TinyDB:
Page 29
TinyDB has been developed at University of Berkeley.
Itsupportsonlyessentialfunctionalitiesforsensor
applications.
Mostofthesensorapplicationsareusedtofilterout
somedatasotheyjustneedtoselectdatawithgiven
conditions.
TinyDBsupportsonlySELECToperationofthe
standardSQL.
Its memory footprint is only 3KBytes.

30. PowerpointTemplates
Comparison:
Page 30
Target Devices Mobile DBMSs
ExtremelySmall
DeviceswithLow
ComputingPower
Sensors TinyDB
Smartcards PicoDBMS
Small Devices with
High Computing
Power
Cell Phones,
PDAs, Car
Navigators,
Ultra Books
Sybase SQL Anywhere,
Oracle Lite,
MS SQL Server CE, SQLite
IBM DB2 Everyplace

31. PowerpointTemplates
Functionalities:
Page 31
TinyDBPicoDBMS
Oracle
Lite
IBM
DB2e
MS SQL Server
Compact
Minimum
Footprint Size
3 KB 30 KB 970 KB320 KB 2 MB
SQL
SELECT
only
a part of
SQL99
a part of
SQL99
a part of
SQL99
a part of SQL99
Views N Y Y Y Y
Integrity
Constraints
N N/A Y Y Y
Concurrency N N Y Y Y
Indexing N Y Y Y Y
Encryption N N/A Y Y Y
Access Control N Y Y Y Y

32. PowerpointTemplates
Supportability of MDBMSRequirements:
Page 32
TinyDBPicoDBMS
Oracle
Lite
IBM
DB2e
MS SQL Server
Compact
Small Footprint Y Y Y Y Y
Flash-Optimized
Storage System
N N N N N
Data
Synchronization
N N Y Y Y
Self-Management Y Y N/A Y N/A
Low Power
Consumption
Y Y N Y Y
Security N Y Y Y Y

33. PowerpointTemplates
References:
Page 33
1.Nori, A. Mobile and embedded databases. In Proc. SIGMOD 2007,
ACM Press (2007), 1175-1177.
2.Tsiftes, N. and Dunkels, A. A database in every sensor. In Proc.
SenSys 2011, ACM Press (2011), 316-332.
3.Whang, K.Y., Song, I.Y., Kim, T.Y., and Lee, K.H. The ubiquitous
DBMS. ACM SIGMOD Record 38, 4 (2009), 14-22.
4.Sen, R. DBMS techniques for lightweight computing devices. In
Proc. MobiDE 2011, ACM Press (2011), 1-8.

34. PowerpointTemplates
ThankYou
Page 34