Implement a secure environment
Each requirement asks one of three questions
Picture one payroll database with three requirements: an Azure-hosted reporting app must sign in without a stored password (who is connecting?), which a managed identity (an identity assigned to the Azure resource that authenticates with it) makes passwordless, with no developer-managed credentials; a stolen backup must be unreadable (where does data become plaintext?), which transparent data encryption (TDE) handles because it encrypts the database, its backups and its transaction log files at rest; and call-center staff may see only an account number's last four digits (which data must be masked in query results?), which dynamic data masking handles because it hides sensitive data in the result sets of queries. Each question matches one of the three areas the study guide lists under this domain, and within an area finer questions separate controls such as TDE and Always Encrypted. The usual mix-up is a control that answers another question: the SQL Server Contributor and SQL DB Contributor roles in Azure role-based access control (Azure RBAC) secure management-related deployment operations, not database connectivity; TDE decrypts each page when it's read into memory, so anyone with the appropriate permissions can read, copy, and share the data; and masking leaves the data in the database unchanged, so it is not encryption. Name the question first: it picks the area, and that area's comparison picks the control.
Authentication and authorization, then data at rest and in transit, then compliance controls
Implement a secure environment carries 20–25% of the DP-300 exam, the same range as two other DP-300 domains, and each of its three areas answers one question. Configure database authentication and authorization answers who is connecting, and what may they do?: authentication is the process of proving the user is who they claim to be, and authorization refers to the permissions that determine what that user is allowed to do, across skills the Microsoft DP-300 study guide lists such as 'Configure Microsoft Entra ID authentication for Azure SQL Database, Azure SQL Managed Instance, and SQL Server', 'Configure security principals' and 'Apply the principle of least privilege for all securables'. Implement security for data at rest and data in transit answers which connections reach the server, and where does data become plaintext?: Microsoft's security overview describes how firewalls prevent network access to the server until you explicitly grant access, and how Transport Layer Security (TLS), TDE and Always Encrypted protect data in transit, at rest and in use, alongside cell-level encryption, which encrypts a column of data by using symmetric encryption in Transact-SQL, under skills the study guide lists such as 'Implement object-level encryption', 'Implement Always Encrypted with secure enclaves' and 'Configure private links and service endpoints' (a virtual network rule references a service endpoint, and Private Link connects through a private endpoint). Implement compliance controls for sensitive data answers which data must be labeled, recorded, masked in query results, limited to certain rows or proven intact?: Data Discovery & Classification discovers, classifies and labels sensitive data, auditing tracks database events, change tracking captures the fact that rows changed while change data capture also captures the changed data, dynamic data masking masks sensitive data to nonprivileged users, row-level security predicates filter the rows available to read operations and block write operations that violate the predicate, and ledger lets you attest that data hasn't been tampered with, under skills the study guide lists such as 'Apply a data classification strategy', 'Implement change data tracking' and 'Implement row-level security'. Microsoft's security overview describes row-level security under authorization and access management, while the study guide lists 'Implement row-level security' under Implement compliance controls for sensitive data, so the domain decision tree asks the third question first, then the second, and a requirement that answers no to both asks the first.
Least privilege across principals, networks and readers
Where two configurations both meet a requirement, least privilege points to the one that grants less. Microsoft's guidance for Azure SQL is to add users to the role with the least privileges required to do their job function and not to assign permissions directly to users, and the study guide states the skill as 'Apply the principle of least privilege for all securables'. The same narrowing carries past permissions. In Azure SQL Database, the Allow Azure services and resources to access this server setting allows connections from all Azure services, including services running in other customers' subscriptions, while virtual network rules accept communications only from selected subnets inside a virtual network. For readers of masked data, the UNMASK permission can be granted at column, table or schema level in SQL Server 2022 and later and in Azure SQL, so original values can go only to the readers who need them.
Three questions and the security controls that answer them
| Question | What it decides | Key controls | Drill into |
|---|---|---|---|
| Who is connecting, and what may they do? | Whether the caller is who it claims to be, then what it is allowed to do | SQL, Windows and Microsoft Entra authentication; logins and database users; database roles and object-level permissions | Configure database authentication and authorization |
| Which connections reach the server, and where does data become plaintext? | Which network origins may connect, and whether data is protected in transit, at rest or in use | IP firewall rules, virtual network rules (for subnets with a service endpoint) and private endpoints (Azure Private Link); TLS; TDE, cell-level encryption, Always Encrypted and Always Encrypted with secure enclaves | Implement security for data at rest and data in transit |
| Which data must be labeled, recorded, masked in query results, limited to certain rows or proven intact? | Which columns are labeled sensitive, which events and changes are recorded, what each reader sees in query results, which rows each user may read or write, and whether tampering can be detected | Data Discovery & Classification, auditing, change tracking, change data capture, dynamic data masking, row-level security and ledger | Implement compliance controls for sensitive data |