Databases Reference
In-Depth Information
CHAPTER 13
Oracle Distributed Databases and
Distributed Data
Data in large and mid-sized companies can sometimes be spread over many different
databases. The data can be on different servers running different operating systems or
even different database management systems. The data needed to answer any specific
business question may need to be accessed from more than one server. A user may need
to access this separate data on several servers simultaneously, or the data required for
an answer may need to be moved to a local server. Inserts, updates, or deletions of data
across these distributed servers may also be necessary.
There are two basic ways to deal with data in distributed databases: as part of a single
distributed entity in which the distributed architecture is transparent, or by using a
variety of replication or data transportation techniques to create copies of the data in
more than one location. This chapter examines these options and the technologies as‐
sociated with each solution.
Of course, there can be performance challenges when there is a need to access data
distributed among multiple databases, sometimes referenced as “federated databases,”
and especially where databases from multiple vendors are mixed. Years of database
query optimization techniques, developed for single databases, must be reproduced
through custom programming to ensure reasonable performance. So, queries across
federated databases are often out of necessity rather than by design. Such queries usually
span databases in a way that was initially considered unlikely, but changing business
needs force the data to be looked at differently. If such queries become a regular and
troublesome occurrence, consolidation of the databases to a single database is likely a
good topic to consider.
