Skip to content

The specialist vocabulary related to database systems can make the subject seem more difficult than it is. The following embedded database terminology list should help lift the veil.


Application Link – Direct link between application program and database functionality. Eliminates client/server overhead of Remote Procedure Calls (RPCs).

Atomicity – A characteristic of transactions where all updates are committed together, or none of them are.

Blob – Binary Large Object. Typically used for un-typed data such as graphics or PDFs.

B-Tree – A page-oriented indexing method that provides fast data access to disk-based contents.

Cache – Server memory containing most recently accessed data.

Concurrency – When multiple computing processes or threads are executed within the same time period.

Connection – The means of communication between a client and a server.

Consistency – Rules enforced by a DBMS to ensure data relationships and contents make sense.

Context – Data associated with a class of operations which allow greater concurrency.

Cost-based Optimization – An optimizing algorithm using database statistics to estimate the best access paths necessary to satisfy a query.

DBMS –Database Management System.

DDL – Database Definition Language.

Distributed Database – Database spread over several computers which have simultaneous access capabilities.

DLL – Dynamic Link Library.

DML – Database Manipulation Language.

Durability – Transactions that will survive system failures.

GUI – Graphical User Interface.

Hash – Indexing method that provides fast data retrieval. Hashing is specifically optimized for keys that have no collating sequence.

Hot Spot – An often-used part of a database that creates a performance bottleneck .

Implicit Locking – Automatic locks that ensure safe operation in a multiuser SQL.

Isolation – In multi-user DBMS usage, the prevention of one user’s uncommitted changes effecting another user.

JDBC – Java Database Connectivity.

Key – A column or combination of columns identified in DDL as identification for rows contained in a table.

Locking – A method for protecting data from being viewed while it may be changed by another user, or from being changed by more than one user at once.

Mirroring – The maintenance of one or more identical copies of an original database.

Multi-platform – Software able to run on many different computers and Operating Systems.

Optimizer – A component of the SQL system that estimates the fastest method to access data.

Primary Key – A Key which provides unique identification of a table row.

Protocol – A specific method in which messages are formulated, formatted, and communicated.

Read-only Transaction – A multi-version concurrency control feature that creates a virtual “snapshot” of a database, viewed by one process without locking  and without seeing another’s modification of that same data.

Record Type – A collection of closely related data fields—equivalent to an SQL table.

Recovery – A procedure following a software or hardware failure which examines a database and all artifacts in order to produce the most recent and reliable state of the data.

Relational Model – A popular database architecture defining tables as a set of rows consisting of sets of columns. The term relational refers to the ability to create relationships (equal, greater-than, …) between column values in different tables resulting in new tables.

Replication – copying data to another database.

Result Set – The complete set of data that is returned by a particular SQL SELECT statement.

Rollback – An operation that cancels changes made by all insert, update and delete statements of a transaction. A rollback may be partial (rolling back changed only to a certain mark in a transaction), or complete (normally called a transaction abort).

Row - One set of related values for all of the columns declared in a given table. Also known as a record occurrence.

Schema – A description of the structure of a database.

Semaphore – Variable used to synchronize shared access to data.

SQL Stored Procedure – A set of statements written in a procedural language containing SQL query and change commands together with variables and conditional control flow.

Synchronization – Multiple methods of preventing independent threads of execution from viewing or modifying common data in an unsafe way.

Table – A collection of closely related data identified as named and typed columns.

Thread – A single, sequential execution of a program segment. Threads can be executed concurrently.

Transaction Log – A record of database changes used to update related databases and facilitate recovery.

 TFS - Transactional File Server.

User-defined Function – An application-specific SQL callable scalar or aggregate function written in C.

User-defined Procedure – An application-specific function written in C and invocable through use of the SQL call statement.

XML – Extensible Markup Language.


Find the full list of database terminology here