Distributed database systems
The users of a database are often widely dispersed, for example, throughout a building,
a city, nationally or internationally. They typically belong to departments or divisions
located at geographically distinct sites. Such a physical organisation necessitates the
linking of sites by a communications network if the users are to access a shared
database. However, transmission speeds across a network are much slower than for
local access to a database, and the extent and complexity of the network make it more
vulnerable to failure. These disadvantages may be alleviated by distributing the
database amongst the user sites. Users can have fast, reliable, local access to data that
they frequently use while still being able to access data that is stored at remote sites.
A distributed database (DDB) is managed by a distributed database management
system (DDBMS) which, as well as providing the usual DBMS functions, keeps track of
how the data is distributed. Consequently, users and applications do not need to know
where the data is stored. To them, the distributed database looks like a single, local
For a distributed database, the flexing methods discussed in the previous chapter can
be extended to include consideration of where data should be located. Different
tables can be located at different sites, the rows and/or columns of a table can be split
between several sites, and copies of the same table (or part thereof) can be stored at
several sites.
The library scenario in section 14.3 assumed that the library operates at only one site.
We now consider an extended scenario in which the service is provided through a
number of branch libraries situated in towns and villages throughout a wider area.
The term global is used to refer to the library as a whole; the term local refers to a
branch library. For example, the global book stock is the total stock held by the library;
the local book stock is the stock held by an individual branch. Each borrower is issued,
on registration, with a plastic card bearing a bar-coded borrowerNo which is globally
unique. Each copy is labelled with a bar-coded accessionNo which is globally unique.
Remember that a distinction was made in section 14.3 between a book (identified by
ISBNX) and a copy (identified by accessionNo). There may be many copies of the same

Get Data Analysis for Database Design, 3rd Edition now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.