4.4. Scalability
Scalability of a database can be interpreted from the perspective
of database architecture and deployment, and also
from the perceptive of the data type (i.e. clinical notes in this
study). Here, we discuss the scalability of the three databases
approaches from the latter perspective. Approach I is outstanding
among the NoSQL approaches concerned in the
paper. Scalability is enabled by allowing additional clinical
concepts to be included by adding new rows to the database
table. However, as each concept occupies a new row in the
table and a general clinical note usually contains a number of
concepts, the size of database will grow exponentially and the
storage space will be consumed significantly, potentially lead
approaches from the latter perspective. Approach I is out
ing to database administration and maintenance problems,
especially for busy clinics or hospitals.
The XML approaches handle clinic notes in a more structured
manner. They are able to handle a large amount of
text data in a single XML field or file while maintaining the
structure of the clinical note. Databases developed with these
approaches can be easily scaled up as there is no limitation on
data granularity or the level of the hierarchy tree required to
represent the clinical notes. Hence, the integrity of data models
can be maintained, where one consultation record always
occupies one row of database entry, regardless of the amountimum.
Modern database supporting XML data type/file can
accept more than 1 gigabyte of data, which is far beyond the
size that a typical textual clinical note would require. Data
storage capacity is therefore not be a limiting factor. This is a
significant advantage of using XML database for clinical data