In June 1970, E. F. Codd of the IBM San Jose Research Laboratory published “A Relational Model of Data for Large Shared Data Banks” in Communications of the ACM. The paper marked the beginning of relational databases as a field, presenting for the first time a complete and mathematically grounded alternative to the database designs then in use.
Before 1970, large databases were built on hierarchical and network models, in which records were linked by physical pointers and programmers had to navigate those links by hand to retrieve data. These systems were fast but rigid: changing the storage structure typically meant rewriting the programs that used it. Codd’s paper argued for protecting users from such internal details through what he called data independence.
By organizing data as relations, simple tables of rows and columns connected by shared values rather than pointers, Codd offered a model that was easier to reason about and that allowed high-level, non-procedural queries. The idea was initially met with skepticism, including from within IBM, partly over whether a relational system could be made to run efficiently.
The publication set off years of research and experimental systems aimed at proving the model practical, and it ultimately reshaped the entire industry. The query languages, design rules, and products that grew out of this single paper became the standard way the world stores and queries structured data.