Chap 1 Introduction to DDBMS

8/4/2019 Chap 1 Introduction to DDBMS

1/34

Introduction to DDBMSIntroduction to DDBMS

What is a distributed DBMS

DBMS and Computer Network

Integration and Centralization

DDBMS environments

Transparency

Advantages


2/34

File SystemsFile Systems

program 1

data description 1

program 2

data description 2

program 3

data description 3

File 1

File 2

File 3


3/34

Database ManagementDatabase Management

database

DBMS

Applicationprogram 1(with datasemantics)



description

manipulation

control


4/34

MotivationMotivation

DatabaseTechnology

ComputerNetworks

integration distribution

integration

integration centralization

Distributed

D

atabaseSystems


5/34

Distributed ComputingDistributed Computing

y A concept in search of a definition and a name.

y A number of autonomous processing elements (notnecessarily homogeneous) that are interconnected by a

computer network and that cooperate in performing

their assigned tasks.


6/34

Distributed ComputingDistributed Computing

y Synonymous terms distributed function

distributed data processing multiprocessors/multicomputers

satellite processing

backend processing

dedicated/special purposecomputers

timeshared systems

functionally modular systems


7/34

What is distributed What is distributed

y Processing logic

y Functions

y Data

y

Control


8/34

What is not a DDBS?What is not a DDBS?

y A timesharing computer system

y A loosely or tightly coupled multiprocessorsystem

y A database system which resides at one of the

nodes of a network of computers - this is acentralized database on a network node


9/34

SharedShared--Memory ArchitectureMemory Architecture

Examples : symmetric multiprocessors (Sequent,Encore) and some mainframes (IBM3090,Bull's DPS8)

1 Pn M

D


10/34

SharedShared--Disk ArchitectureDisk Architecture

Examples : DEC's VAXcluster, IBM's IMS/VSData Sharing

D

P1

M1

Pn

Mn


11/34

SharedShared--Nothing ArchitectureNothing Architecture

Examples : Teradata's DBC, Tandem, Intel'sParagon, NCR's 3600 and 3700

P1

M1

D1

Pn

Mn

Dn


12/34

Centralized DBMS on a NetworkCentralized DBMS on a Network

Site 5

Site 1

Site 2

Site 3Site 4

Communication

Network


13/34

Distributed DBMS EnvironmentDistributed DBMS Environment

Site 5

Site 1

Site 2

Site 3Site 4

Communication

Network


14/34

What is a Distributed Database System?What is a Distributed Database System?

A distributed database (DDB) is a collection ofmultiple, logically interrelateddatabases distributed overa computer network.

A distributed database management system (DDBMS)is the software that manages the DDB and provides anaccess mechanism that makes this distribution

transparent to the users.

Distributed database system (DDBS) = DDB + DDBMS


15/34

Implicit AssumptionsImplicit Assumptions

y Data stored at a number of sites each sitelogicallyconsists of a single processor.

y Processors at different sites areinterconnected by a computer network nomultiprocessors parallel database systems

y Distributed database is a database, not acollection of files data logically related as

exhibited in the users access patterns relational data model

y D-DBMS is a full-fledged DBMS not remote file system


16/34

ApplicationsApplications

y Manufacturing - especially multi-plantmanufacturing

y Military command and controly EFTy Corporate MISy Airlinesy

Hotel chainsy Any organization which has a

decentralized organization structure


17/34

Distributed DBMS PromisesDistributed DBMS Promises

Transparent management of distributed,fragmented, and replicated data

Improved reliability/availability throughdistributed transactions

Improved performance

Easier and more economical systemexpansion


18/34

TransparencyTransparency

y Transparency is the separation of the higherlevel semantics of a system from the lowerlevel implementation issues.

y Fundamental issue is to provide

data independencein the distributed environment

Network (distribution) transparency

Replication transparency

Fragmentation transparency

x horizontal fragmentation: selectionx vertical fragmentation: projection

xhybrid


19/34

ExampleExample

TITLE SAL

PAY

Elect. Eng. 40000

Syst. Anal. 34000

Mech. Eng. 27000

Programmer 24000

PROJ

PNO PNAME BUDGET

ENO ENAME TITLE

E1 J. Doe Elect. Eng.E2 M. Smith Syst. Anal.E3 A. Lee Mech. Eng.

E4 J. Miller Programmer

E5B.

Casey Syst. Anal.E6 L. Chu Elect. Eng.

E7 R. Davis Mech. Eng.

E8 J. Jones Syst. Anal.

EMP

ENO PNO RESP

E1 P1 Manager 12

DUR

E2 P1 Analyst 24E2 P2 Analyst 6E3 P3 Consultant 10E

3P

4E

ngineer 48

E4 P2 Programmer 18E5 P2 Manager 24E6 P4 Manager 48

E7 P3 Engineer 36

E8 P3 Manager 40

ASG

P1 Instrumentation 150000

P3 CAD/C AM 250000

P2 Database Develop. 135000

P4 Maintenance 310000

E7 P5 Engineer 23


20/34

Transparent AccessTransparent Access

SELECTENAME,SAL

FROM EMP,ASG,PAY

WHERE DUR > 12

AND EMP.ENO = ASG.ENOAND PAY.TITLE = EMP.TITLE Paris projects

Paris employees

Paris assignments

Boston employees

Montreal projects

Paris projects

Montreal employees

Montreal assignments

Boston

Communication

Network

Montreal

Paris

New

York

Boston projects

Boston employees

Boston assignments

Boston projects

New York employees

New York projects

New York assignments

Tokyo


21/34

Distributed DatabaseDistributed Database -- User ViewUser View

Distributed Database


22/34

Distributed DBMSDistributed DBMS -- RealityReality

CommunicationSubsystem

UserQuery

DBMSSoftware

DBMSSoftware

UserApplication

DBMSSoftware

UserApplicationUser

QueryDBMS

Software

User

Query

DBMS

Sof

tware


23/34

Potentially Improved PerformancePotentially Improved Performancey A distributed DBMS fragments the conceptual db enabling

data to be stored in close proximity to its points of use(localization)

Each site handles only a portion of db so contention for CPUand IO services is not serious

Localization reduces remote access delays

y Inherent parallelism of distributed system may be exploitedfor :

Inter query parallelism

Intra query parallelism


24/34

Inter and Intra Query ParallelismInter and Intra Query Parallelism

y Inter Query Parallelism Ability to

execute multiple queries at the same time

y Intra Query Parallelism Ability to breakup a single query into a number of sub

query each of which is executed at a

different site, accessing a different part ofthe distributed database


25/34

Parallelism RequirementsParallelism Requirements

y If user access to distributed db consist ofonly read only access then much of thedatabase should be replicated

y If the db access are not only read only buta mix of read and update operations then the database will requireimplementation of concurrency controland commit protocols


26/34

Complicating FactorsComplicating Factors

y Complexity of the architecture

y

Cost of implementation

y Distribution of control

y Security


27/34

Distributed DBMSDistributed DBMS Problem AreasProblem Areas

yDistributed Database Design

how to distribute the database

replicated & non-replicated database distribution

yQuery Processing

Design algorithms that convert user transactionsto data manipulation instructions

optimization problem

min{cost = data transmission + local processing}


28/34

y Distributed Directory Management

Directory contains information about data items inthe database

Address the problem of how to arrange the directory

global to entire DDBMS or local to each site Single copy / Multiple copy

y Operating System Support operating system with proper support for database

operations

dichotomy between general purpose processingrequirements and database processing requirements



29/34


y Concurrency Control

Integration of the database is maintained

Most extensive studied problem in DDBMS

Mutual consistency All values of multiple copiesof every data item to converge to the same value

Achieved through locking or time stamping

y Reliability

how to make the system resilient to failures

consistency


30/34

Directory

Management

Relationship Between IssuesRelationship Between Issues

Reliability

Deadlock

Management

Query

Processing

ConcurrencyControl

Distribution

Design


31/34

ExerciseExercise

y A bank has 3 branches at different

location. At each branch a computercontrols the teller terminals of the branch

and the account database of that branch.y Please draw two distributed database high

level designs for the same, one for LAN

and the other for WAN


32/34

Distributed database on LANDistributed database on LAN


33/34

Distributed database on WANDistributed database on WAN


34/34

Multiprocessor SystemMultiprocessor System

Documents

Chap 1 Introduction to DDBMS