Databases Reference
In-Depth Information
Figure 5-3
Memory
Scaling Up
…
Scaling Out
Scale up/scale out.
taken into account during database design. For example, different DBMS systems
support different options for how data is stored and organized within the data-
base tables. Microsoft SQL Server, for example, supports a data type that directly
supports storage and manipulation of XML data.
Data types
are data storage for-
mats used when defining database tables. Most other DBMSs, at least at this point
in their development, do not include a data type designed for this purpose. XML
data storage requires you to develop your own work-arounds based on more tra-
ditional data types.
Certain hardware characteristics, such as processor speeds and disk data
transfer rates, are associated with the physical database design process though
not directly part of it. Simply put, the faster the hardware, the more tolerant the
system can be of a less than optimal physical design. In fact, some database
designers might try to scale up or scale out the hardware platform rather than
correcting problems found in the physical design after the database has been
deployed. Options are shown in Figure 5-3.
Scaling up
refers to improving the
server hardware, such as by installing additional processors, more memory, or
faster disk subsystems.
Scaling out
refers to spreading the data across multiple
database servers in a distributed data environment. Both are valid responses to
changing conditions such as increased user support requirements or an increase
in the volume of data stored on the database. However, these responses can also
be used in attempt to cover up an inefficient design.
