There down approach will start from scratch and the

 

There are different methods to
design the Distributed Database like Top Down and Bottom Up approaches. The top
down approach will start from scratch and the second bottom up approach will
the multi database or existing database. in this paper we indicate how to build
the requirements about the distribution of data and application and how to
build the distribution of a schemes.

Introductions:

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

In this paper we will Design the
Distributed Database, it must be two thing to design the distributed DBMS the
distributed DBMS software  and the
distribution of application program that run on it. The DBMS concern  the two key things Schema Design (Similar
to  Centralized Database Design), and
Schema Distribution . There are two also key design strategies to design the
distributed database system, the 1st one is top town approach and the other one
is bottom up approach the top down approach will start from scratch and the
second bottom up approach will the multi database or existing database. The top
town approach the designer easily understand the requirements of the users and
they can easily to relocate into formal requirement. The requirement document
input of two activities view design and the conceptual design. 2 In the top
down approach, the designer performs abstract, logical, and physical style
phases, that increasingly refine high-level, system-independent specifications
of the data base in to low-level, system-dependent specifications, and designer
ignore any detail regarding the physical implementation of the distributed
database. The bottom-up approach assumes, instead, that a specification of the
data

 

 

bases at each site exists
already, either because there are a unit existing databases that have to be unified
to form a multi-database or as a results of the abstract specification of the
databases has been in deep irritate each web site severally. In either case,
the positioning specifications got to be integrated in order to return up with
a worldwide specification. While top-down and bottom-up approaches appear to
represent two extremes, in many smart cases the designer yield half bottom-up
and half top down. we’ll present the two approaches in separate sections
therefore discuss the interaction between them. The user doesn’t be want to
aware  that the where come from it is
locally or distributed DBMS. But one thing keeping in your mind that the
distributed design is to achieve maximum locality of data and applications. “90
percent of the data should be found at the local site, and only 10percent of
the data should be accessed on a remote site”1. The time and the cost
make an important role in DDB.

Distributed Database System

 

 

 

 

 

 

 

 

 

There are some steps to design the
distributed database system.

·       
Analysis of 
the external, application requirements

·       
Design of the global schema

·       
Design of the fragmentation

·       
Design of the distribution schema

·       
Design of the local schemes

·       
Design of the local physical layers

 

Top-down Approach:

Top Down Design Process 2

 

A general technique for coming up
with centralized databases includes four phases: needs analysis, abstract
style, logical style, and physical design, needs Analysis deals with the
gathering of users’ unstructured specifications of the information application,
associated produces an unambiguous definition and classification of the weather
to be thought-about within the style of the information. the knowledge} is
collected in an exceedingly style data wordbook. abstract style, typically
additional divided visible  style and
think about Integration, produces a abstract specification of a world,
integrated information schema and of the applications that ar performed
thereon. native style transforms the integrated abstract schema into a
information schema of a given package sort (Relational, Network,or
Hierarchical).The choice of package sort are suffering from the wants of the
abstract model additionally as by pragmatic issues. Physical style is performed
in line with the capabilities and options of the actual package chosen, and
produces the definition of physical access structures that implement the
information. the planning of distribution adds to the on top of phases a
further one, referred to as Distribution style, that assumes as input a world,
website-independent schema and produces as a result the sub schema for every
site of the distributed information. in essence, distribution style may be
applied to any of the worldwide abstract, logical, or physical schema. This
alternative is subject to the subsequent trade off: Details concerning
implementation ought to be set only the information distribution is given, to
permit concentrating on the physical style of every native information
severally. freelance physical style is necessary if the location package are
heterogeneous. On the opposite hand, a particular description of information
and operations helps in estimating the performance of varied distributions.
This trade off suggests that the distribution style ought to be performed at
the start of the logical style part. At that point, information and operations
ar delineate exactly and also the 1st implementation issues ar thought-about.

 

 

 

 

Design of the Fragmentation:

 

 

Entire table is not a suitable unit,
so dividing a table into smaller ones. There are two
alternatives for
dividing a table
1. Vertically
2. Horizontally

The purpose of this phase is to
determine the non-overlapping pieces, fragments of the global database which
can be stored as a

unit on different sites. The data
elements having the same properties, behavior are assigned to the same
fragment. Main aspects of the fragmentation:

Granularity:

The granularity determines at
which level of database storage can be the fragmentation performed. If it is
too low (field) then it needs

a lot of management cost. If it
is too rough (user level) then the unnecessary elements should be replicated
causing a higher cost.

Fragmentation strategy

– horizontal fragmentation:. In
this case the granularity is at the tuple level, and the attribute values of
the tuple determine the corresponding fragment. The relation is partitioned
horizontally

into fragments.

– vertical fragmentation: the
assignment of the data elements in a table is based on the schema position of
the data. In this case the different projections are the content of the
fragments.

– mixed fragmentation: the data
are fragmented by both the vertical as the horizontal method.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

                                   

 

                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                           

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1)    
STEFAN0 CERI, BARBARA PERNICI,
AND G I 0 WIEDERHOLD
‘Distributed
Database Design Methodologies’ 1987

2)    
Yao et al., 1982a