Databases Reference
In-Depth Information
same, but then the
Ordered
set associated to that product can be used to find and
evaluate the orders of other customers. In this case, only customers having actually
ordered that product are considered.
To achieve a similar behavior in the hierarchical model, IBM IMS provides bidi-
rectional logical relationships [
97
] that are supported by a subset of its hierarchical
database types. With these, logical paths from order line nodes of one physical
hierarchy can be modeled that point to nodes of ordered products stored in an
another physical hierarchy and from the product nodes back to the first hierarchy to
all order line nodes that contain this product. In contrast to the network model, this
approach needs additional logical structures and pointers, increasing the overhead
for the DBMS.
Relational Model
In the late 1970s, IBM developed “System R” as an experimental prototype for
a DBMS that uses the relational database model (RDBMS - relational database
management system). According to Astrahan et al. [
6
] System R was intended
to demonstrate that a relational system can be used in real environments with a
comparable performance to the existing systems of that time. The structured query
language (SQL), initial called “SEQUEL” (structured English query language) [
27
],
was developed as a query language to retrieve and manipulate stored data in
System R. To test and evaluate System R, several installations at internal locations of
IBM were set up. The project ended in 1979 with the result that the relational model
is the basis for a viable database technology with commercial potential [
93
,p.85].
Chamberlin et al. [
28
] came to the conclusion that databases on the basis of the
relational model are able to support many concurrent users, who perform repetitive
transactions and ad hoc queries. They argue that the high-level user interface enabled
by the relational database model positively effects user productivity for developing
new applications.
Along with System R, INGRES (Interactive Graphics and Retrieval System) was
one of the first projects to provide a proof of concept for a practical DBMS based on
the relational model. INGRES was started as a research project at the University
of Berkeley, California. The motivation behind INGRES was the utilization of
two basic characteristics of the relational model, namely, the high degree of data
independence and the possibility for a high level query language [
199
]. INGRES
included an own language called “QUEry Language” (QUEL). When SQL evolved
as the standard database language [
50
], INGRES was converted to supporting
SQL [
93
, p. 86], while continuing to support QUEL.
IBM's System R and INGRES, however, were not the first to spawn a commercial
relational database for the market. Greenwald et al. [
77
] claim that Oracle V.2,
released in 1979, was the worlds first commercial relational database. According
to Hernandez and Viescas [
93
], the first commercial version of INGRES entered
the market in 1981. IBM announced its own RDBMS called SQL/Data System in
1981 [
41
] and shipped it starting in 1982. In 1983, IBM announced Database 2
Search WWH ::
Custom Search