Microsoft SQL Server (often shortened to MS SQL) is a relational database management system (RDBMS) built by Microsoft. It stores data in tables, is queried with Transact-SQL (T-SQL) — Microsoft's dialect of SQL — and scales from a single desktop application to enterprise systems serving thousands of concurrent users.
What is Microsoft SQL Server?
SQL Server is a relational database management system designed to store, query and manage data reliably. Data lives in tables made of rows and columns, and relationships between tables are expressed with primary key and foreign key constraints. The engine guarantees that transactions follow the ACID properties: atomicity, consistency, isolation and durability.
SQL Server is more than a storage engine. It ships with server-side programming tools such as stored procedures, triggers, views, user-defined functions, indexes and constraints. On top of that, it offers separate service components for reporting, data integration and multidimensional analysis.
A short history: from Sybase roots to today
SQL Server's roots go back to the late 1980s and a collaboration between Microsoft, Sybase and Ashton-Tate. The first release was based on Sybase's database engine and targeted the OS/2 operating system. After Microsoft and Sybase parted ways in the mid-1990s, Microsoft continued to develop the product on its own for Windows NT. This shared origin explains why SQL Server and Sybase (SAP ASE) still have similarities, such as the Transact-SQL language.
Over the years the product matured with a query optimizer, high-availability features, columnar storage, in-memory processing and cloud integration. Releases are named after their year, for example SQL Server 2016, 2017, 2019 and 2022.
Architecture and core components
SQL Server is made up of several services that can be installed independently. A server may run only the database engine, or you can add analysis and reporting components as your needs grow.
| Component | Abbreviation | Role |
|---|---|---|
| Database Engine | — | Storage, querying, transactions and security; the core of the system |
| SQL Server Agent | — | Runs scheduled jobs (backups, maintenance, ETL) and raises alerts |
| Integration Services | SSIS | Extract, transform and load (ETL) workflows |
| Analysis Services | SSAS | Multidimensional (OLAP) and tabular analytical models |
| Reporting Services | SSRS | Server-based report authoring and delivery |
| Full-Text Search | — | Word- and phrase-based search over text columns |
Inside the database engine you'll find subsystems such as the query optimizer that plans how a query runs, the storage engine that moves data between disk and memory, and the log manager that writes the transaction log.
A single machine can run more than one SQL Server instance: one default instance and several named instances. Each instance has its own settings and user databases. Every instance also ships with a few system databases the engine needs to operate:
- master: holds server-level configuration and a record of every database.
- model: acts as the template for each newly created database.
- msdb: stores SQL Server Agent jobs, schedules and backup history.
- tempdb: a workspace for temporary tables, sorting and row versions that is recreated every time the server restarts.
Transact-SQL (T-SQL): the query language
SQL Server uses Transact-SQL, which is standard SQL plus the procedural extensions Microsoft added on top. T-SQL provides programming constructs such as variables, flow control (IF, WHILE), error handling (TRY...CATCH) and stored procedures.
-- Create a table and run a sample query
CREATE TABLE Articles (
Id INT IDENTITY(1,1) PRIMARY KEY,
Title NVARCHAR(200) NOT NULL,
Published DATE NOT NULL,
Views INT NOT NULL DEFAULT 0
);
INSERT INTO Articles (Title, Published)
VALUES (N'What Is SQL Server?', '2026-09-28');
SELECT Title, Views
FROM Articles
WHERE Published >= '2026-01-01'
ORDER BY Views DESC;Stored procedures let you keep repeated logic on the server. This reduces the number of round trips over the network and makes it easier to grant permissions at the procedure level.
CREATE PROCEDURE dbo.IncrementViews
@ArticleId INT
AS
BEGIN
SET NOCOUNT ON;
UPDATE Articles
SET Views = Views + 1
WHERE Id = @ArticleId;
END;Editions and deployment options
SQL Server is not a single product but a family of editions aimed at different needs. Editions differ in supported features, scale limits and licensing.
| Edition | Intended use | Note |
|---|---|---|
| Enterprise | Large, mission-critical workloads | The broadest feature set and scalability |
| Standard | Mid-sized applications | Core database features with capped resources |
| Web | Hosting / web scenarios | Licensing terms suited to web hosters |
| Developer | Development and testing | Same features as Enterprise, free; not for production |
| Express | Small apps, learning | Free; comes with database size and resource limits |
Windows, Linux and container support
SQL Server was originally a Windows-only product. That changed with SQL Server 2017, which added support for Linux, making it installable on distributions such as Red Hat Enterprise Linux, Ubuntu and SUSE. Official Docker images also allow it to run in containers and on Kubernetes.
High availability, backup and recovery
For systems expected to stay online, SQL Server offers several high-availability (HA) and disaster-recovery (DR) options:
- Always On Availability Groups: keep a group of databases on multiple replicas, synchronously or asynchronously, with automatic or manual failover when the primary goes down.
- Backup and restore: a flexible recovery strategy built from full, differential and transaction-log backups.
- Recovery models: Simple, Full and Bulk-logged determine how the transaction log is kept, based on how much data loss you can tolerate.
- Replication: distributing or copying data across multiple servers.
Security features
SQL Server supports two authentication modes: Windows authentication only, and mixed mode (Windows plus SQL Server logins). Authorization is managed through server- and database-level roles and object-level permissions. Key data-protection capabilities include:
- Transparent Data Encryption (TDE): encrypts database files and backups at rest.
- Always Encrypted: encrypts sensitive columns on the client side, so keys are never exposed to the server in plain form.
- Row-Level Security (RLS): row-level access control so a user only sees the rows they are authorized to see.
- Dynamic Data Masking: hides sensitive fields (such as a card number) behind masks in query results.
Performance: indexing, columnstore and in-memory
Good indexing is the foundation of query performance. SQL Server distinguishes between a clustered index, which physically orders the table, and a nonclustered index, which provides a separate lookup structure. On top of these it offers columnstore indexes for analytical workloads and In-Memory OLTP for tables kept in memory.
For the optimizer to make good decisions, the statistics that summarize how data is distributed need to stay current. To track how query plans change over time and catch regressions, the Query Store feature keeps historical plan and runtime data inside the database.
On the concurrency side, snapshot-based isolation levels can reduce cases where locks block readers; in this approach row versions are stored in tempdb. Choosing the right isolation level for your application sets the balance between consistency and concurrency.
SQL Server in the cloud: the Azure SQL family
Microsoft also offers the SQL Server engine as managed cloud services, which hand a large part of infrastructure maintenance to the provider:
- Azure SQL Database: a fully managed, single-database platform service (PaaS).
- Azure SQL Managed Instance: a managed instance model with high compatibility with on-premises SQL Server.
- SQL Server on Azure VM: a traditional, fully controlled install on a virtual machine (IaaS).
For a broader comparison of managed database offerings, see our guide on cloud database services.
When should you choose SQL Server?
SQL Server is a natural fit for teams that want strong transactional guarantees, a mature tooling ecosystem and tight integration with the Microsoft stack (Windows Server, .NET, Azure, Active Directory). Having enterprise reporting, ETL and analysis components under one roof is an advantage in many scenarios.
That said, SQL Server is just one relational option among many. If open source and licensing cost come first, consider PostgreSQL or MySQL; if you want a large-scale enterprise alternative with a long track record, look at Oracle Database. To see the options as a whole, our database types guide is a good starting point.
Management and development tools
The SQL Server ecosystem comes with a mature toolset. The most widely used are SQL Server Management Studio (SSMS) for graphical administration, Azure Data Studio as a lightweight cross-platform alternative, sqlcmd for the command line, and bcp for bulk data transfer.
For developers, there are official drivers such as ADO.NET, Entity Framework, JDBC, ODBC and OLE DB, along with client libraries for many languages. That means SQL Server is comfortable to use not only from .NET but also from stacks like Java, Python or Node.js. For monitoring health, built-in diagnostics such as Dynamic Management Views (DMVs) and Extended Events are available.
A short note on licensing
The paid editions of SQL Server are generally offered under two licensing models: per-core licensing and the Server + Client Access License (Server + CAL) model. Which model fits depends on the edition, the number of cores, and the number of users or devices.