Databases have become ubiquitous, to the extent that the term is used casually and broadly to refer to any collection of data. We describe the initial motivations that lead to database technology in distinction to simple files. Many associated requirements, as ease of use, long-term availability, continuous operation, and an ability to recover from errors, have become standard features. Early inventions include schemas (1958), hierarchical structures (1963), general Object-attribute-value structures (1969), B-tree indexes (1972), and view mechanisms (1975).
A formal definition of database operations, the relational model, due to Ted Codd, of IBM, published in 1970, initiated academic interest. By 1985 databases based on that model became the de-facto standard, to the extent that today the terms have become synonymous. Alternate database models have been proposed and implemented, but have not had a wide overall impact. Some recent database proposals mimic, unwittingly, earlier implementations, and show similar benefits and problems. But some may serve applications that are now poorly served, and where some losses are offset by better specific performance or lower costs.
The growth of hardware capabilities, specifically memory and storage sizes have shifted the range where comprehensive databases are needed, and enable alternate designs as well. The major problem for new systems is the need to support all the expectations that existing commercial implementations generate.
An open question is the role of open database systems in the future. We will leave time to address questions.