Database Reference
In-Depth Information
major enterprise dollars. The relational database market wasn't just a boon for propri-
etary software vendors. Many popular open-source relational databases were devel-
oped, including the popular projects MySQL and PostgreSQL. Although Codd never
attained the massive financial success of Ellison, his work was so inf luential it earned
him a Turing Award (sometimes called the computer science Nobel Prize) in 1981.
The Relational Database ACID Test
As relational databases were brought to market, Codd was not completely satisfied
with the design of various implementations. He went on to publish a set of rules that
he believed must be met in order for a database to qualify as “relational.” Some of
these rules described standards for data consistency, including the need for all informa-
tion in a relational database (including table and column names) to be represented in
only one way. Other rules, which are still debated today, include the manner in which
a relational database should represent NULL values.
The former Royal Air Force pilot's rules for relational databases were never com-
pletely realized by most mainstream software packages. Whether a piece of software
fulfilled all of Codd's rules or simply most of them, relational databases all shared
similar utility and f lexibility that have helped them become the dominant manner of
interacting with digital data. Databases that generally conform to a relational model
are described as exhibiting four traits known as
ACID
—an acronym representing the
terms atomicity, consistency, isolation, and durability. In summary, these four rules
attempt to prevent unexpected or erratic results.
Atomicity
and
consistency
refer to
relational databases' goals of preventing partial insertions and to maintaining a valid
state from one transaction to another.
Isolation
refers to the ability of a database to
handle multiple transactions at the same time without conflicts.
Durability
means
that a database retains the last known correct state in the event of errors or crashes.
When launching a piece of software, technological limitations and decisions about
acceptable compromises outweigh the goal of conforming to a particular idealized
architecture. Codd publicly voiced a distaste for products that billed themselves as rela-
tional but did not conform to his strict idealization of the relational concept. Codd's
tension would foreshadow an even larger deviation of database architecture from the
ideal relational model in the decades to follow.
One of the many reasons that the relational database concept became the dominant
form of data management is that it strives to provide a great deal of data consistency.
Surely a database administrator should be wary of transactions that leave the database
in an uncertain state. Despite these consistency features, the relational model has been
disrupted by the same humble invention that disrupted so many other aspects of daily
life: the Internet. All of a sudden, application development was shifting to the Web.
Applications of all types, from business software to email to video, were being built
Search WWH ::
Custom Search