Database Reference
In-Depth Information
user need not be concerned while coding his or her query. This distributed data inde-
pendence is, in fact, a natural extension of the principles of logical and physical data
independence.
Pursue further the consideration of a transaction executed from a certain loca-
tion. The transaction may operate on pieces of data stored at different locations. If
the transaction completes successfully and commits, then all changes to data by the
transaction must persist; if the transaction aborts, none of the changes must remain
in the database. In other words, the atomicity of the transaction in the distributed
environment must be ensured as if it executed in a centralized database system.
Let us define a distributed database system. We will explore a few fundamental
concepts and then review the goals of distributed database systems. That will lead
into a review of the advantages and disadvantages.
What is a Distributed Database?
Let us begin with a standard definition:
A distributed database
is a related collection of shared data of an organization along with the description of
that data, logically integrated and interrelated, by physically distributed across loca-
tions over a computer network.
Figure 18-1 shows an illustration of a distributed database.
A distributed computing system consists of a number of processing units inter-
connected by a communications network. These processing units need not be homo-
geneous, but they have to cooperate with one another in processing queries and
London
Paris
New York
Tokyo
Hong Kong
Figure 18-1
Distributed database.
Search WWH ::




Custom Search