Microsoft Azure SQL Managed Instance Security Technical Implementation Guide

Within the database, object ownership implies full privileges to the owned object, including the privilege to assign access to the owned objects to other subjects. Database functions and procedures can be coded using definer's rights. This allows anyone who utilizes the object to perform the actions if they were the owner. If not properly managed, this can lead to privileged actions being taken by unauthorized individuals.

If the Azure SQL Managed Instance were to allow any user to make changes to database structure or logic, then those changes might be implemented without undergoing the appropriate testing and approvals that are part of a robust change management process. Accordingly, only qualified and authorized individuals will be allowed to obtain access to information system components for purposes of initiating changes, including upgrades and modifications. Unmanaged changes that occur to the database software libraries or configuration can lead to unauthorized or compromised installations.

Applications, including DBMSs, must prevent unauthorized and unintended information transfer via shared system resources. Data used for the development and testing of applications often involves copying data from production. It is important that specific procedures exist for this process, to include the conditions under which such transfer may take place, where the copies may reside, and the rules for ensuring sensitive data are not exposed. Copies of sensitive data must not be misplaced or left in a temporary location without the proper controls.

With respect to database management systems, one class of threat is known as SQL Injection, or more generally, code injection. It takes advantage of the dynamic execution capabilities of various programming languages, including dialects of SQL. In such cases, the attacker deduces the manner in which SQL statements are being processed, either from inside knowledge or by observing system behavior in response to invalid inputs. When the attacker identifies scenarios where SQL queries are being assembled by application code (which may be within the database or separate from it) and executed dynamically, the attacker is then able to craft input strings that subvert the intent of the query. Potentially, the attacker can gain unauthorized access to data, including security settings, and severely corrupt or destroy the database. The principal protection against code injection is not to use dynamic execution except where it provides necessary functionality that cannot be utilized otherwise. Use strongly typed data items rather than general-purpose strings as input parameters to task-specific, precompiled stored procedures and functions (and triggers). When dynamic execution is necessary, ways to mitigate the risk include the following, which must be implemented both in the on-screen application and at the database level, in the stored procedures: - Allow strings as input only when necessary. - Rely on data typing to validate numbers, dates, etc. Do not accept invalid values. If substituting other values for them, think carefully about whether this could be subverted. - Limit the size of input strings to what is truly necessary. - If single quotes/apostrophes, double quotes, semicolons, equal signs, angle brackets, or square brackets will never be valid as input, reject them. - If comment markers will never be valid as input, reject them. In SQL, these are -- or /* */ . - If HTML and XML tags, entities, comments, etc., will never be valid, reject them. - If wildcards are present, reject them unless truly necessary. In SQL these are the underscore and the percentage sign, and the word "ESCAPE" is also a clue that wildcards are in use. - If SQL key words, such as SELECT, INSERT, UPDATE, DELETE, CREATE, ALTER, DROP, ESCAPE, UNION, GRANT, REVOKE, DENY, MODIFY will never be valid, reject them. Use case-insensitive comparisons when searching for these. Bear in mind that some of these words, particularly "Grant" (as a person's name), could also be valid input. - If there are range limits on the values that may be entered, enforce those limits. - Institute procedures for inspection of programs for correct use of dynamic coding, by a party other than the developer. - Conduct rigorous testing of program modules that use dynamic coding, searching for ways to subvert the intended use. - Record the inspection and testing in the system documentation. - Bear in mind that all this applies not only to screen input, but also to the values in an incoming message to a web service or to a stored procedure called by a software component that has not itself been hardened in these ways. Not only can the caller be subject to such vulnerabilities; it may itself be the attacker. This calls for inspection of application source code, which will require collaboration with the application developers. It is recognized that in many cases, the database administrator (DBA) is organizationally separate from the application developers, and may have limited, if any, access to source code. Nevertheless, protections of this type are so important to the secure operation of databases that they must not be ignored. At a minimum, the DBA must attempt to obtain assurances from the development organization that this issue has been addressed and must document what has been discovered. Satisfies: SRG-APP-000251-DB-000392, SRG-APP-000251-DB-000391

Without the association of security labels to information, there is no basis for Azure SQL Managed Instance to make security-related access-control decisions. Security labels are abstractions representing the basic properties or characteristics of an entity (e.g., subjects and objects) with respect to safeguarding information. These labels are typically associated with internal data structures (e.g., tables, rows) within the database and are used to enable the implementation of access control and flow control policies; reflect special dissemination, handling, or distribution instructions; or support other aspects of the information security policy. One example includes marking data as classified or CUI. These security labels may be assigned manually or during data processing, but, either way, it is imperative these assignments are maintained. If the security labels are lost, there is the risk of a data compromise. The mechanism used to support security labeling may be a feature of Azure SQL Managed Instance, a third-party product, or custom application code. Satisfies: SRG-APP-000313-DB-000309, SRG-APP-000311-DB-000308, SRG-APP-000314-DB-000310

Discretionary Access Control (DAC) is based on the notion that individual users are "owners" of objects and therefore have discretion over who should be authorized to access the object and in which mode (e.g., read or write). Ownership is usually acquired as a consequence of creating the object or via specified ownership assignment. DAC allows the owner to determine who will have access to objects they control. An example of DAC includes user-controlled table permissions. When discretionary access control policies are implemented, subjects are not constrained with regard to what actions they can take with information for which they have already been granted access. Thus, subjects that have been granted access to information are not prevented from passing (i.e., the subjects have the discretion to pass) the information to other subjects or objects. A subject that is constrained in its operation by Mandatory Access Control policies is still able to operate under the less rigorous constraints of this requirement. Thus, while Mandatory Access Control imposes constraints preventing a subject from passing information to another subject operating at a different sensitivity level, this requirement permits the subject to pass the information to any subject at the same sensitivity level. The policy is bound by the information system boundary. Once the information is passed outside of the control of the information system, additional means may be required to ensure the constraints remain in effect. While the older, more traditional definitions of discretionary access control require identity-based access control, that limitation is not required for this use of discretionary access control.

In certain situations, to provide required functionality, a database management system (DBMS) needs to execute internal logic (stored procedures, functions, triggers, etc.) and/or external code modules with elevated privileges. However, if the privileges required for execution are at a higher level than the privileges assigned to organizational users invoking the functionality applications/programs, those users are indirectly provided with greater privileges than assigned by organizations. Privilege elevation by "execute as" must be utilized only where necessary and protected from misuse.

Allowing regular users to install software, without explicit privileges, creates the risk that untested or potentially malicious software will be installed on the system. Explicit privileges (escalated or administrative privileges) provide the regular user with explicit capabilities and control that exceed the rights of a regular user. Azure SQL Managed Instance functionality and the nature and requirements of databases will vary, so while users are not permitted to install unapproved software, there may be instances where the organization allows the user to install approved software packages such as from an approved software repository. The requirements for production servers will be more restrictive than those used for development and research. Azure SQL Managed Instance must enforce software installation by users based upon what types of software installations are permitted (e.g., updates and security patches to existing software) and what types of installations are prohibited (e.g., software whose pedigree with regard to being potentially malicious is unknown or suspect) by the organization). In the case of a database management system, this requirement covers stored procedures, functions, triggers, views, etc.

Failure to provide logical access restrictions associated with changes to configuration may have significant effects on the overall security of the system. When dealing with access restrictions pertaining to change control, it should be noted that any changes to the hardware, software, and/or firmware components of the information system can potentially have significant effects on the overall security of the system. Accordingly, only qualified and authorized individuals must be allowed to obtain access to system components for the purposes of initiating changes, including upgrades and modifications.

Azure SQL Managed Instance databases handling data requiring data-at-rest protections must employ cryptographic mechanisms to prevent unauthorized disclosure and modification of the information at rest. Selection of a cryptographic mechanism is based on the need to protect the integrity of organizational information. The strength of the mechanism is commensurate with the security category and/or classification of the information. The decision whether and what to encrypt rests with the data owner and is also influenced by the physical measures taken to secure the equipment and media on which the information resides.

Azure SQL Managed Instance handling data requiring data-at-rest protections must employ cryptographic mechanisms to prevent unauthorized disclosure and modification of the information at rest. Selection of a cryptographic mechanism is based on the need to protect the integrity of organizational information. The strength of the mechanism is commensurate with the security category and/or classification of the information.

A common vulnerability is unplanned behavior when invalid inputs are received. This requirement guards against adverse or unintended system behavior caused by invalid inputs, where information system responses to the invalid input may be disruptive or cause the system to fail into an unsafe state. The behavior will be derived from the organizational and system requirements and includes, but is not limited to, notification of the appropriate personnel, creating an audit record, and rejecting invalid input. This calls for inspection of application source code, which will require collaboration with the application developers. It is recognized that in many cases, the database administrator (DBA) is organizationally separate from the application developers, and may have limited, if any, access to source code. Nevertheless, protections of this type are so important to the secure operation of databases that they must not be ignored. At a minimum, the DBA must attempt to obtain assurances from the development organization that this issue has been addressed and must document what has been discovered.

Nonrepudiation of actions taken is required to maintain data integrity. Examples of particular actions taken by individuals include creating information, sending a message, approving information (e.g., indicating concurrence or signing a contract), and receiving a message. Nonrepudiation protects against later claims by a user of not having created, modified, or deleted a particular data item or collection of data in the database. In designing a database, the organization must define the types of data and the user actions that must be protected from repudiation. The implementation must then include building audit features into the application data tables and configuring Azure SQL Managed Instance's audit tools to capture the necessary audit trail. Design and implementation also must ensure that applications pass individual user identification to Azure SQL Managed Instance, even where the application connects to Azure SQL Managed Instance with a standard, shared account. Satisfies: SRG-APP-000080-DB-000063, SRG-APP-000815-DB-000160

Nonrepudiation of actions taken is required to maintain data integrity. Examples of particular actions taken by individuals include creating information, sending a message, approving information (e.g., indicating concurrence or signing a contract), and receiving a message. Nonrepudiation protects against later claims by a user of not having created, modified, or deleted a particular data item or collection of data in the database. In designing a database, the organization must define the types of data and the user actions that must be protected from repudiation. The implementation must then include building audit features into the application data tables and configuring the DBMS's audit tools to capture the necessary audit trail. Design and implementation also must ensure that applications pass individual user identification to the DBMS, even where the application connects to the DBMS with a standard, shared account.

Under some circumstances, it may be useful to monitor who/what is reading privilege/permission/role information; therefore, it must be possible to configure auditing to do this. DBMSs typically make such information available through views or functions. This requirement addresses explicit requests for privilege/permission/role membership information. It does not refer to the implicit retrieval of privileges/permissions/role memberships that Azure SQL Managed Instance continually performs to determine if any and every action on the database is permitted. To aid in diagnosis, it is necessary to keep track of failed attempts in addition to the successful ones. Satisfies: SRG-APP-000091-DB-000066, SRG-APP-000091-DB-000325, SRG-APP-000495-DB-000326

Session auditing is used when a user's activities are under investigation. To ensure capture of all activity during those periods when session auditing is in use, it must be in operation for the whole time Azure SQL Managed Instance is running.

Information systems are capable of providing a wide variety of functions and services. Some of the functions and services, provided by default, may not be necessary to support essential organizational operations (e.g., key missions, functions). It is detrimental for software products to provide, or install by default, functionality exceeding requirements or mission objectives. Examples include, but are not limited to, installing advertising software, demonstrations, or browser plugins not related to requirements or providing a wide array of functionality, not required for every mission, that cannot be disabled. Azure SQL Managed Instance must adhere to the principles of least functionality by providing only essential capabilities. Demonstration and sample database objects and applications present publicly known attack points for malicious users. These demonstration and sample objects are meant to provide simple examples of coding specific functions and are not developed to prevent vulnerabilities from being introduced to the Azure SQL Managed Instance.

To prevent unauthorized connection of devices, unauthorized transfer of information, or unauthorized tunneling (i.e., embedding of data types within data types), organizations must disable or restrict unused or unnecessary physical and logical ports/protocols/services on information systems. Applications are capable of providing a wide variety of functions and services. Some of the functions and services provided by default may not be necessary to support essential organizational operations. Additionally, it is sometimes convenient to provide multiple services from a single component (e.g., email and web services); however, doing so increases risk over limiting the services provided by any one component. To support the requirements and principles of least functionality, the application must support the organizational requirements providing only essential capabilities and limiting the use of ports, protocols, and/or services to only those required, authorized, and approved to conduct official business or to address authorized quality of life issues. Database management systems using ports, protocols, and services deemed unsafe are open to attack through those ports, protocols, and services. This can allow unauthorized access to the audit and through the database to other components of the information system.

To prevent unauthorized connection of devices, unauthorized transfer of information, or unauthorized tunneling (i.e., embedding of data types within data types), organizations must disable or restrict unused or unnecessary physical and logical ports/protocols/services on information systems. Applications are capable of providing a wide variety of functions and services. Some of the functions and services provided by default may not be necessary to support essential organizational operations. Additionally, it is sometimes convenient to provide multiple services from a single component (e.g., email and web services); however, doing so increases risk over limiting the services provided by any one component. To support the requirements and principles of least functionality, the application must support the organizational requirements providing only essential capabilities and limiting the use of ports, protocols, and/or services to only those required, authorized, and approved to conduct official business or to address authorized quality of life issues. Database management systems using ports, protocols, and services deemed unsafe are open to attack through those ports, protocols, and services. This can allow unauthorized access to the database and through the database to other components of the information system. Satisfies: SRG-APP-000142-DB-000094, SRG-APP-000383-DB-000364

To ensure accountability and prevent unauthenticated access, organizational users must be identified and authenticated to prevent potential misuse and compromise of the system. Organizational users include organizational employees or individuals the organization deems to have equivalent status of employees (e.g., contractors). Organizational users (and any processes acting on behalf of users) must be uniquely identified and authenticated for all accesses, except the following: (i) Accesses explicitly identified and documented by the organization. Organizations document specific user actions that can be performed on the information system without identification or authentication; and (ii) Accesses that occur through authorized use of group authenticators without individual authentication. Organizations may require unique identification of individuals using shared accounts, for detailed accountability of individual activity.

The DOD standard for authentication is DOD-approved PKI certificates. Once a PKI certificate has been validated, it must be mapped to an Azure SQL Managed Instance database user account or login for the authenticated identity to be meaningful to Azure SQL Managed Instance and useful for authorization decisions.

Nonorganizational users include all information system users other than organizational users, which include organizational employees or individuals the organization deems to have equivalent status of employees (e.g., contractors, guest researchers, individuals from allied nations). Nonorganizational users must be uniquely identified and authenticated for all accesses other than those accesses explicitly identified and documented by the organization when related to the use of anonymous access, such as accessing a web server. Accordingly, a risk assessment is used in determining the authentication needs of the organization. Scalability, practicality, and security are simultaneously considered in balancing the need to ensure ease of use for access to federal information and information systems with the need to protect and adequately mitigate risk to organizational operations, assets, individuals, and other organizations.

Information system management functionality includes functions necessary to administer databases, network components, workstations, or servers and typically requires privileged user access. The separation of user functionality from information system management functionality is either physical or logical and is accomplished by using different computers, different central processing units, different instances of the operating system, different network addresses, combinations of these methods, or other methods, as appropriate. An example of this type of separation is observed in web administrative interfaces that use separate authentication methods for users of any other information system resources. This may include isolating the administrative interface on a different domain and with additional access controls. If administrative functionality or information regarding Azure SQL Managed Instance management is presented on an interface available for users, information on Azure SQL Managed Instance settings may be inadvertently made available to the user.

To ensure sufficient storage capacity for the audit logs, the database management system (DBMS) must be able to allocate audit record storage capacity. Although another requirement (SRG-APP-000515-DB-000318) mandates that audit data be off-loaded to a centralized log management system, it remains necessary to provide space on the database server to serve as a buffer against outages and capacity limits of the off-loading mechanism. The task of allocating audit record storage capacity is usually performed during initial installation of the DBMS and is closely associated with the database administrator (DBA) and system administrator (SA) roles. The DBA or SA will usually coordinate the allocation of physical drive space with the application owner/installer and the application will prompt the installer to provide the capacity information, the physical location of the disk, or both. In determining the capacity requirements, consider such factors as: total number of users; expected number of concurrent users during busy periods; number and type of events being monitored; types and amounts of data being captured; the frequency/speed with which audit records are off-loaded to the central log management system; and any limitations that exist on the DBMS's ability to reuse the space formerly occupied by off-loaded records.

Auditing for Azure SQL Managed Instance tracks database events and writes them to an audit log in the Azure storage account, Log Analytics workspace, or Event Hubs. Under normal conditions, the audit space allocated by an Azure Storage account can grow quite large. Since a requirement exists to halt processing upon audit failure, a service outage would result.

Changes in the database objects (tables, views, procedures, functions) that record and control permissions, privileges, and roles granted to users and roles must be tracked. Without an audit trail, unauthorized changes to the security subsystem could go undetected. The database could be severely compromised or rendered inoperative. Satisfies: SRG-APP-000496-DB-000334, SRG-APP-000496-DB-000335

Changes in categories of information must be tracked. Without an audit trail, unauthorized access to protected data could go undetected. To aid in diagnosis, it is necessary to keep track of failed attempts in addition to the successful ones. For detailed information on categorizing information, refer to FIPS Publication 199, Standards for Security Categorization of Federal Information and Information Systems, and FIPS Publication 200, Minimum Security Requirements for Federal Information and Information Systems. Satisfies: SRG-APP-000498-DB-000347, SRG-APP-000498-DB-000346

Information system auditing capability is critical for accurate forensic analysis. Reconstruction of harmful events or forensic analysis is not possible if audit records do not contain enough information. To support analysis, some types of events will need information to be logged that exceeds the basic requirements of event type, time stamps, location, source, outcome, and user identity. If additional information is not available, it could negatively impact forensic investigations into user actions or other malicious events. The organization must determine what additional information is required for complete analysis of the audited events. The additional information required is dependent on the type of information (e.g., sensitivity of the data and the environment within which it resides). At a minimum, the organization must employ either full-text recording of privileged commands or the individual identities of users of shared accounts, or both. The organization must maintain audit trails in sufficient detail to reconstruct events to determine the cause and impact of compromise. Examples of detailed information the organization may require in audit records are full-text recording of privileged commands or the individual identities of shared account users.

The removal of security objects from the database/database management system (DBMS) would seriously degrade a system's information assurance posture. If such an action is attempted, it must be logged. To aid in diagnosis, it is necessary to keep track of failed attempts in addition to the successful ones. Satisfies: SRG-APP-000501-DB-000337, SRG-APP-000501-DB-000336

Changes in categorized information must be tracked. Without an audit trail, unauthorized access to protected data could go undetected. To aid in diagnosis, it is necessary to keep track of failed attempts in addition to the successful ones. For detailed information on categorizing information, refer to FIPS Publication 199, Standards for Security Categorization of Federal Information and Information Systems, and FIPS Publication 200, Minimum Security Requirements for Federal Information and Information Systems. Satisfies: SRG-APP-000502-DB-000349, SRG-APP-000502-DB-000348

For completeness of forensic analysis, it is necessary to track failed attempts to log on to Azure SQL Managed Instance. While positive identification may not be possible in a case of failed authentication, as much information as possible about the incident must be captured. Satisfies: SRG-APP-000503-DB-000351, SRG-APP-000503-DB-000350

Without tracking privileged activity, it would be difficult to establish, correlate, and investigate the events relating to an incident or identify those responsible for one. System documentation should include a definition of the functionality considered privileged. A privileged function in this context is any operation that modifies the structure of the database, its built-in logic, or its security settings. This would include all Data Definition Language (DDL) statements and all security-related statements. In an Azure SQL Managed Instance environment, it encompasses but is not necessarily limited to: CREATE ALTER DROP GRANT REVOKE DENY There may also be Data Manipulation Language (DML) statements that, subject to context, should be regarded as privileged. Possible examples in SQL Managed Instance include: TRUNCATE TABLE; DELETE, or DELETE affecting more than n rows, for some n, or DELETE without a WHERE clause; UPDATE or UPDATE affecting more than n rows, for some n, or UPDATE without a WHERE clause; any SELECT, INSERT, UPDATE, or DELETE to an application-defined security table executed by other than a security principal. Depending on the capabilities of Azure SQL Managed Instance and the design of the database and associated applications, audit logging may be achieved by means of DBMS auditing features, database triggers, other mechanisms, or a combination of these. Note that it is particularly important to audit, and tightly control, any action that weakens the implementation of this requirement itself, since the objective is to have a complete audit trail of all administrative activity. Satisfies: SRG-APP-000504-DB-000354, SRG-APP-000504-DB-000355

For completeness of forensic analysis, it is necessary to know how long a user's (or other principal's) connection to the Azure SQL Managed Instance lasts. This can be achieved by recording disconnections, in addition to logons/connections, in the audit logs. Disconnection may be initiated by the user or forced by the system (as in a timeout) or result from a system or network failure. To the greatest extent possible, all disconnections must be logged.

For completeness of forensic analysis, it is necessary to track who logs on to Azure SQL Managed Instance. Concurrent connections by the same user from multiple workstations may be valid use of the system; or such connections may be due to improper circumvention of the requirement to use the CAC for authentication; or they may indicate unauthorized account sharing; or they may be because an account has been compromised. If the fact of multiple, concurrent logons by a given user can be reliably reconstructed from the log entries for other events (logons/connections; voluntary and involuntary disconnections), it is not mandatory to create additional log entries specifically for this.

Without tracking all or selected types of access to all or selected objects (tables, views, procedures, functions, etc.), it would be difficult to establish, correlate, and investigate the events relating to an incident, or identify those responsible for one. In an Azure SQL Managed Instance environment, types of access include, but are not necessarily limited to: SELECT INSERT UPDATE DELETE EXECUTE Satisfies: SRG-APP-000507-DB-000356, SRG-APP-000507-DB-000357

In this context, direct access is any query, command, or call to Azure SQL Managed Instance that comes from any source other than the application(s) it supports. For example, the command line or a database management utility program. The intent is to capture all activity from administrative and nonstandard sources.

Information stored in one location is vulnerable to accidental or incidental deletion or alteration. When configured and enabled, Azure SQL Managed Instance only supports writing audit records to a container within an Azure Blob Storage Account. To mitigate the risk associated with maintaining a single copy of this audit data, the blob container used to store the audit data generated by this Azure SQL Managed Instance must be protected by an immutable policy. This policy must be configured to lock blobs for an organizationally defined period of time.

Automated mechanisms for centralized reviews and analyses include Security Information and Event Management products.

Audit information includes all information needed to successfully audit system activity, such as audit records, audit log settings, audit reports, and personally identifiable information. Audit logging tools are those programs and devices used to conduct system audit and logging activities. Protection of audit information focuses on technical protection and limits the ability to access and execute audit logging tools to authorized individuals. Physical protection of audit information is addressed by both media protection controls and physical and environmental protection controls. The common language runtime (CLR) component of the .NET Framework for Microsoft Windows in SQL allows the user to write stored procedures, triggers, user-defined types, user-defined functions, user-defined aggregates, and streaming table-valued functions, using any .NET Framework language, including Microsoft Visual Basic .NET and Microsoft Visual C#. CLR packing assemblies can access resources protected by .NET Code Access Security when it runs managed code. Specifying UNSAFE enables the code in the assembly complete freedom to perform operations in the SQL process space that can potentially compromise the robustness of SQL. UNSAFE assemblies can also potentially subvert the security system of either SQL or the common language runtime.

The database management system (DBMS) must prevent the installation of organization-defined software and firmware components without verification that the component has been digitally signed using a certificate recognized and approved by the organization. The common language runtime (CLR) component of the .NET Framework for Microsoft Windows in SQL allows the user to write stored procedures, triggers, user-defined types, user-defined functions, user-defined aggregates, and streaming table-valued functions, using any .NET Framework language, including Microsoft Visual Basic .NET and Microsoft Visual C#. CLR packing assemblies can access resources protected by .NET Code Access Security when it runs managed code. Specifying UNSAFE enables the code in the assembly complete freedom to perform operations in the SQL process space that can potentially compromise the robustness of SQL. UNSAFE assemblies can also potentially subvert the security system of either SQL or the common language runtime.

Password-based authentication applies to passwords regardless of whether they are used in single-factor or multifactor authentication. Long passwords or passphrases are preferable over shorter passwords. Enforced composition rules provide marginal security benefits while decreasing usability. However, organizations may choose to establish certain rules for password generation (e.g., minimum character length for long passwords) under certain circumstances and can enforce this requirement in IA-5(1)(h). Account recovery can occur, for example, in situations when a password is forgotten. Cryptographically protected passwords include salted one-way cryptographic hashes of passwords. The list of commonly used, compromised, or expected passwords includes passwords obtained from previous breach corpuses, dictionary words, and repetitive or sequential characters. The list includes context-specific words, such as the name of the service, username, and derivatives thereof.

If the system were to allow any user to make changes to software libraries, then those changes might be implemented without undergoing the appropriate testing and approvals that are part of a robust change management process. Accordingly, only qualified and authorized individuals will be allowed to obtain access to information system components for purposes of initiating changes, including upgrades and modifications. Unmanaged changes that occur to the database software libraries or configuration can lead to unauthorized or compromised installations.

If the system were to allow any user to make changes to software libraries, then those changes might be implemented without undergoing the appropriate testing and approvals that are part of a robust change management process. Accordingly, only qualified and authorized individuals will be allowed to obtain access to information system components for purposes of initiating changes, including upgrades and modifications. Unmanaged changes that occur to the database software libraries or configuration can lead to unauthorized or compromised installations.

Weak passwords may be easily guessed. When passwords are used to encrypt keys used for encryption of sensitive data, then the confidentiality of all data encrypted using that key is at risk.

When not encrypted by the SMK, system administrators or application administrators may access and use the DMK to view sensitive data that they are not authorized to view. Where alternate encryption means are not feasible, encryption by the SMK may be necessary. To help protect sensitive data from unauthorized access by DBAs, mitigations may be in order. Mitigations may include automatic alerts or other audit events when the DMK is accessed outside of the application or by a DBA account.

Backup and recovery of the Certificate used for encryption is critical to the complete recovery of the database. Not having this key can lead to loss of data during recovery.

Invalid user input occurs when a user inserts data or characters into an application's data entry fields and the application is unprepared to process that data. This results in unanticipated application behavior, potentially leading to an application or information system compromise. Invalid user input is one of the primary methods employed when attempting to compromise an application. With respect to database management systems, one class of threat is known as SQL Injection, or more generally, code injection. It takes advantage of the dynamic execution capabilities of various programming languages, including dialects of SQL. Potentially, the attacker can gain unauthorized access to data, including security settings, and severely corrupt or destroy the database. Even when no such hijacking takes place, invalid input that gets recorded in the database, whether accidental or malicious, reduces the reliability and usability of the system. Available protections include data types, referential constraints, uniqueness constraints, range checking, and application-specific logic. Application-specific logic can be implemented within the database in stored procedures and triggers, where appropriate. This calls for inspection of application source code, which will require collaboration with the application developers. It is recognized that in many cases, the database administrator (DBA) is organizationally separate from the application developers, and may have limited, if any, access to source code. Nevertheless, protections of this type are so important to the secure operation of databases that they must not be ignored. At a minimum, the DBA must attempt to obtain assurances from the development organization that this issue has been addressed and must document what has been discovered.

Nonrepudiation of actions taken is required to maintain data integrity. Examples of particular actions taken by individuals include creating information, sending a message, approving information (e.g., indicating concurrence or signing a contract), and receiving a message. Nonrepudiation protects against later claims by a user of not having created, modified, or deleted a particular data item or collection of data in the database. Azure SQL Managed Instance provides the ability for high privileged accounts to impersonate users in a database using the TRUSTWORTHY feature. This will allow members of the fixed database role to impersonate any user within the database.

Without the capability to generate audit records, it would be difficult to establish, correlate, and investigate the events relating to an incident or identify those responsible for one. Audit records can be generated from various components within Azure SQL Managed Instance (e.g., process, module). Certain specific application functionalities may be audited as well. The list of audited events is the set of events for which audits are to be generated. This set of events is typically a subset of the list of all events for which the system is capable of generating audit records. DOD has defined the list of events for which Azure SQL Managed Instance will provide an audit record generation capability as the following: (i) Successful and unsuccessful attempts to access, modify, or delete privileges, security objects, security levels, or categories of information (e.g., classification levels); (ii) Access actions, such as successful and unsuccessful logon attempts, privileged activities, or other system-level access, starting and ending time for user access to the system, concurrent logons from different workstations, successful and unsuccessful accesses to objects, all program initiations, and all direct access to the information system; and (iii) All account creation, modification, disabling, and termination actions. Organizations may define additional events requiring continuous or ad hoc auditing.

Without the capability to restrict which roles and individuals can select which events are audited, unauthorized personnel may be able to prevent or interfere with the auditing of critical events. Suppression of auditing could permit an adversary to evade detection. Misconfigured audits can degrade the system's performance by overwhelming the audit log. Misconfigured audits may also make it more difficult to establish, correlate, and investigate the events relating to an incident or identify those responsible for one.

If audit data were to become compromised, then competent forensic analysis and discovery of the true source of potentially malicious system activity is difficult, if not impossible, to achieve. In addition, access to audit records provides information an attacker could potentially use to their advantage. To ensure the veracity of audit data, the information system and/or the application must protect audit information from any and all unauthorized access. This includes read, write, copy, etc. This requirement can be achieved through multiple methods, which will depend upon system architecture and design. Some commonly employed methods include ensuring log files enjoy the proper file system permissions utilizing file system protections and limiting log data location. Additionally, applications with user interfaces to audit records must not allow the unfettered manipulation of or access to those records via the application. If the application provides access to the audit data, the application becomes accountable for ensuring that audit information is protected from unauthorized access. Azure SQL Managed Instance is able to view and manipulate audit file data. Audit information includes all information (e.g., audit records, audit settings, and audit reports) needed to successfully audit information system activity. Satisfies: SRG-APP-000118-DB-000059, SRG-APP-000119-DB-000060, SRG-APP-000120-DB-000061

Protecting audit data also includes identifying and protecting the tools used to view and manipulate log data; therefore, protecting audit tools is necessary to prevent unauthorized operation on audit data. Applications providing tools to interface with audit data will leverage user permissions and roles identifying the user accessing the tools and the corresponding rights the user enjoys to make access decisions regarding the deletion of audit tools. Audit tools include, but are not limited to, vendor-provided and open source audit tools needed to successfully view and manipulate audit information system activity and records. Audit tools include custom queries and report generators. This focuses on external tools for log maintenance and review. Other STIG requirements govern SQL privileges to maintain trace or audit definitions. Satisfies: SRG-APP-000121-DB-000202, SRG-APP-000122-DB-000203, SRG-APP-000123-DB-000204

Information systems are capable of providing a wide variety of functions and services. Some of the functions and services, provided by default, may not be necessary to support essential organizational operations (e.g., key missions, functions). It is detrimental for applications to provide, or install by default, functionality exceeding requirements or mission objectives. Applications must adhere to the principles of least functionality by providing only essential capabilities. Azure SQL Managed Instance may spawn additional external processes to execute procedures that are defined in the Azure SQL Managed Instance but stored in external host files (external procedures). The spawned process used to execute the external procedure may operate within a different OS security context than Azure SQL Managed Instance and provide unauthorized access to the host system. The xp_cmdshell extended stored procedure allows execution of host executables outside the controls of database access permissions. This access may be exploited by malicious users who have compromised the integrity of the Azure SQL Managed Instance database process to control the host operating system to perpetrate additional malicious activity.

Information systems are capable of providing a wide variety of functions and services. Some of the functions and services, provided by default, may not be necessary to support essential organizational operations (e.g., key missions, functions). It is detrimental for applications to provide, or install by default, functionality exceeding requirements or mission objectives. Applications must adhere to the principles of least functionality by providing only essential capabilities. Azure SQL Managed Instance may spawn additional external processes to execute procedures that are defined in the Azure SQL Managed Instance but stored in external host files (external procedures). The spawned process used to execute the external procedure may operate within a different OS security context than Azure SQL Managed Instance and provide unauthorized access to the host system. The common language runtime (CLR) component of the .NET Framework for Microsoft Windows in Azure SQL Managed Instance allows a user to write stored procedures, triggers, user-defined types, user-defined functions, user-defined aggregates, and streaming table-valued functions, using any .NET Framework language, including Microsoft Visual Basic .NET and Microsoft Visual C#. CLR packing assemblies can access resources protected by .NET Code Access Security when it runs managed code. Specifying UNSAFE enables the code in the assembly complete freedom to perform operations in the Azure SQL Managed Instance process space that can potentially compromise the robustness of Azure SQL Managed Instance. UNSAFE assemblies can also potentially subvert the security system of either Azure SQL Managed Instance or the common language runtime.

Information systems are capable of providing a wide variety of functions and services. Some of the functions and services, provided by default, may not be necessary to support essential organizational operations (e.g., key missions, functions). It is detrimental for applications to provide, or install by default, functionality exceeding requirements or mission objectives. Applications must adhere to the principles of least functionality by providing only essential capabilities. Azure SQL Managed Instance may spawn additional external processes to execute procedures that are defined in the Azure SQL Managed Instance but stored in external host files (external procedures). The spawned process used to execute the external procedure may operate within a different OS security context than Azure SQL Managed Instance and provide unauthorized access to the host system. A linked server allows for access to distributed, heterogeneous queries against OLE DB data sources. After a linked server is created, distributed queries can be run against this server, and queries can join tables from more than one data source. If the linked server is defined as an instance of SQL Server, Azure SQL Database, or Azure SQL Managed Instance, remote stored procedures can be executed. This access may be exploited by malicious users who have compromised the integrity of the Azure SQL Managed Instance.

OS/enterprise authentication and identification must be used (SRG-APP-000023-DB-000001). Native DBMS authentication may be used only when circumstances make it unavoidable; and must be documented and AO-approved. The DOD standard for authentication is DOD-approved PKI certificates. Authentication based on User ID and Password may be used only when it is not possible to employ a PKI certificate and requires AO approval. In such cases, the DOD standards for password complexity and lifetime must be implemented. DBMS products that can inherit the rules for these from the operating system or access control program (e.g., Microsoft Active Directory or Microsoft Entra) must be configured to do so. For other DBMSs, the rules must be enforced using available configuration parameters or custom code.

If Azure SQL Managed Instance provides too much information in error logs and administrative messages to the screen, this could lead to compromise. The structure and content of error messages must be carefully considered by the organization and development team. The extent to which the information system is able to identify and handle error conditions is guided by organizational policy and operational requirements. Some default database management system (DBMS) error messages can contain information that could aid an attacker in, among other things, identifying the database type, host address, or state of the database. Custom errors may contain sensitive customer information. It is important that detailed error messages be visible only to those who are authorized to view them; that general users receive only generalized acknowledgment that errors have occurred; and that these generalized messages appear only when relevant to the user's task. For example, a message along the lines of, "An error has occurred. Unable to save your changes. If this problem persists, please contact your help desk." would be relevant. A message such as "Warning: your transaction generated a large number of page splits" would likely not be relevant. "ABGQ is not a valid widget code." would be appropriate; but "The INSERT statement conflicted with the FOREIGN KEY constraint 'WidgetTransactionFK'. The conflict occurred in database 'DB7', table 'dbo.WidgetMaster', column 'WidgetCode''' would not, as it reveals too much about the database structure.

Preventing nonprivileged users from executing privileged functions mitigates the risk that unauthorized individuals or processes may gain unnecessary access to information or privileges. System documentation should include a definition of the functionality considered privileged. Depending on circumstances, privileged functions can include, for example, establishing accounts, performing system integrity checks, or administering cryptographic key management activities. Nonprivileged users are individuals that do not possess appropriate authorizations. Circumventing intrusion detection and prevention mechanisms or malicious code protection mechanisms are examples of privileged functions that require protection from nonprivileged users. A privileged function in Azure SQL Managed Instance/database context is any operation that modifies the structure of the database, its built-in logic, or its security settings. This would include all Data Definition Language (DDL) statements and all security-related statements. In an Azure SQL Managed Instance environment, it encompasses but is not necessarily limited to: CREATE ALTER DROP GRANT REVOKE DENY There may also be Data Manipulation Language (DML) statements that, subject to context, should be regarded as privileged. Possible examples include: TRUNCATE TABLE; DELETE, or DELETE affecting more than n rows, for some n, or DELETE without a WHERE clause; UPDATE or UPDATE affecting more than n rows, for some n, or UPDATE without a WHERE clause; Any SELECT, INSERT, UPDATE, or DELETE to an application-defined security table executed by other than a security principal. Depending on the capabilities of Azure SQL Managed Instance and the design of the database and associated applications, the prevention of unauthorized use of privileged functions may be achieved by means of database management system (DBMS) security features, database triggers, other mechanisms, or a combination of these.

Failure to provide logical access restrictions associated with changes to configuration may have significant effects on the overall security of the system. When dealing with access restrictions pertaining to change control, it must be noted that any changes to the hardware, software, and/or firmware components of the information system can potentially have significant effects on the overall security of the system. Accordingly, only qualified and authorized individuals must be allowed to obtain access to system components for the purposes of initiating changes, including upgrades and modifications.

Failure to provide logical access restrictions associated with changes to configuration may have significant effects on the overall security of the system. When dealing with access restrictions pertaining to change control, it must be noted that any changes to the hardware, software, and/or firmware components of the information system can potentially have significant effects on the overall security of the system. Accordingly, only qualified and authorized individuals must be allowed to obtain access to system components for the purposes of initiating changes, including upgrades and modifications.

Without auditing the enforcement of access restrictions against changes to configuration, it would be difficult to identify attempted attacks and an audit trail would not be available for forensic investigation for after-the-fact actions. Enforcement actions are the methods or mechanisms used to prevent unauthorized changes to configuration settings. Enforcement action methods may be as simple as denying access to a file based on the application of file permissions (access restriction). Audit items may consist of lists of actions blocked by access restrictions or changes identified after the fact.

Database management systems can maintain separate execution domains for each executing process by assigning each process a separate address space. Each process has a distinct address space so that communication between processes is controlled through the security functions, and one process cannot modify the executing code of another process. Maintaining separate execution domains for executing processes can be achieved, for example, by implementing separate address spaces.

Changes to the security configuration must be tracked. This requirement applies to situations where security data is retrieved or modified via data manipulation operations, as opposed to via specialized security functionality. In an SQL environment, types of access include, but are not necessarily limited to: - SELECT - INSERT - UPDATE - DELETE - EXECUTE To aid in diagnosis, it is necessary to keep track of failed attempts in addition to the successful ones. Satisfies: SRG-APP-000492-DB-000332, SRG-APP-000492-DB-000333

Changes in categorized information must be tracked. Without an audit trail, unauthorized access to protected data could go undetected. For detailed information on categorizing information, refer to FIPS Publication 199, Standards for Security Categorization of Federal Information and Information Systems, and FIPS Publication 200, Minimum Security Requirements for Federal Information and Information Systems. To aid in diagnosis, it is necessary to keep track of failed attempts in addition to the successful attempts. Satisfies: SRG-APP-000494-DB-000345, SRG-APP-000494-DB-000344

Changes in the permissions, privileges, and roles granted to users and roles must be tracked. Without an audit trail, unauthorized elevation or restriction of privileges could go undetected. Elevated privileges give users access to information and functionality that they should not have; restricted privileges wrongly deny access to authorized users. In an Azure SQL Managed Instance environment, adding permissions is typically done via the GRANT command, or, in the negative, the DENY command. To aid in diagnosis, it is necessary to keep track of failed attempts in addition to the successful ones.

Changes in the permissions, privileges, and roles granted to users and roles must be tracked. Without an audit trail, unauthorized elevation or restriction of privileges could go undetected. Elevated privileges give users access to information and functionality that they should not have; restricted privileges wrongly deny access to authorized users. In an SQL environment, modifying permissions is typically done via the GRANT, REVOKE, and DENY commands. To aid in diagnosis, it is necessary to keep track of failed attempts in addition to the successful ones. Satisfies: SRG-APP-000495-DB-000329, SRG-APP-000495-DB-000328

Changes in the permissions, privileges, and roles granted to users and roles must be tracked. Without an audit trail, unauthorized elevation or restriction of privileges could go undetected. Elevated privileges give users access to information and functionality that they should not have; restricted privileges wrongly deny access to authorized users. In an Azure SQL Managed Instance environment, deleting permissions is typically done via the REVOKE or DENY command. To aid in diagnosis, it is necessary to keep track of failed attempts in addition to the successful ones. Satisfies: SRG-APP-000499-DB-000330, SRG-APP-000499-DB-000331

Azure SQL Managed Instance [sa] account has special privileges required to administer the database. The [sa] account is a well-known account and is likely to be targeted by attackers and thus more prone to providing unauthorized access to the database. This [sa] default account is administrative and could lead to catastrophic consequences, including the complete loss of control over Azure SQL Managed Instance. If the [sa] default account is not disabled, an attacker could gain access through the account. Azure SQL Managed Instance by default disables the [sa] account at creation. Some applications that run on Azure SQL Managed Instance require the [sa] account to be enabled for the application to function properly. These applications that require the [sa] account to be enabled are usually legacy systems.

Azure SQL Managed Instance's [sa] account has special privileges required to administer the database. The [sa] account is a well-known account name that is likely to be targeted by attackers and is thus, more prone to providing unauthorized access to the database. Since [sa] is administrative in nature, the compromise of a default account can have catastrophic consequences, including the complete loss of control over Azure SQL Managed Instance. Since SQL Server needs this account to exist and it must not be removed, one way to mitigate this risk is to change the [sa] account name.

Information systems are capable of providing a wide variety of functions and services. Some of the functions and services, provided by default, may not be necessary to support essential organizational operations (e.g., key missions, functions). It is detrimental for applications to provide, or install by default, functionality exceeding requirements or mission objectives. Applications must adhere to the principles of least functionality by providing only essential capabilities. Azure SQL Managed Instance may spawn additional external processes to execute procedures that are defined in Azure SQL Managed Instance but stored in external host files (external procedures). The spawned process used to execute the external procedure may operate within a different OS security context than Azure SQL Managed Instance and provide unauthorized access to the host system. SQL Server is capable of providing a wide range of features and services. Some of the features and services, provided by default, may not be necessary, and enabling them could adversely affect the security of the system. The Allow Filesystem Enumeration feature allows access to the file system, potentially providing unauthorized access and must be disabled on secure systems.

Information systems are capable of providing a wide variety of functions and services. Some of the functions and services, provided by default, may not be necessary to support essential organizational operations (e.g., key missions, functions). It is detrimental for applications to provide, or install by default, functionality exceeding requirements or mission objectives. Applications must adhere to the principles of least functionality by providing only essential capabilities. Azure SQL Managed Instance may spawn additional external processes to execute procedures that are defined in Azure SQL Managed Instance but stored in external host files (external procedures). The spawned process used to execute the external procedure may operate within a different OS security context than Azure SQL Managed Instance and provide unauthorized access to the host system. Azure SQL Managed Instance is capable of providing a wide range of features and services. CLR Strict Security interprets assemblies as unsafe, ignoring permissions on individual assemblies which otherwise may be able to access external system resources, call unmanaged code, and acquire sysadmin privileges.

Information systems are capable of providing a wide variety of functions and services. Some of the functions and services, provided by default, may not be necessary to support essential organizational operations (e.g., key missions, functions). It is detrimental for applications to provide, or install by default, functionality exceeding requirements or mission objectives. Applications must adhere to the principles of least functionality by providing only essential capabilities. Azure SQL Managed Instance may spawn additional external processes to execute procedures that are defined in the Azure SQL Managed Instance but stored in external host files (external procedures). The spawned process used to execute the external procedure may operate within a different OS security context than Azure SQL Managed Instance and provide unauthorized access to the host system. Azure SQL Managed Instance is capable of providing a wide range of features and services. Some of the features and services, provided by default, may not be necessary, and enabling them could adversely affect the security of the system. The Hadoop Connectivity feature allows multiple types of external data sources to be created and used across all sessions on the server. An exploit to the Azure SQL Managed Instance could result in a compromise of the host system and external Azure SQL Managed Instance resources.

Azure SQL Managed Instance is capable of providing a wide range of features and services. Some of the features and services, provided by default, may not be necessary, and enabling them could adversely affect the security of the system. Enabling Replication XPs opens a significant attack surface area that can be used by an attacker to gather information about the system and potentially abuse the privileges of Azure SQL Managed Instance.

To prevent the compromise of authentication information, such as passwords and PINs, during the authentication process, the feedback from the information system must not provide any information that would allow an unauthorized user to compromise the authentication mechanism. Obfuscation of user-provided information when typed into the system is a method used in addressing this risk. For example, displaying asterisks when a user types in a password or PIN, is an example of obscuring feedback of authentication information. Database applications may allow for entry of the account name and password as a visible parameter of the application execution command. This practice must be prohibited and disabled to prevent shoulder surfing.