Understand security policies

Any code that a report server executes must be part of a specific code access security policy. These security policies consist of code groups that map evidence to a set of named permission sets. Often, code groups are associated with a named permission set that specifies the allowable permissions for code in that group. The runtime uses evidence provided by a trusted host or by the loader to determine which code groups the code belongs to and, therefore, which permissions to grant the code. Reporting Services adheres to this security policy architecture as defined by the Microsoft .NET Framework common language runtime (CLR). The following sections describe the various types of code in Reporting Services and the policy rules associated with them.

Report server assemblies

Report server assemblies contain code that is part of the Reporting Services product. Reporting Services is written using managed code assemblies; all of these assemblies are strong-named (that is, digitally signed). The code groups for these assemblies are defined using the StrongNameMembershipCondition, which provides evidence based on public key information for the assembly's strong name. The code group is granted the FullTrust permission set.

Report server extensions (rendering, data, delivery, and security)

Report server extensions are custom data, delivery, rendering, and security extensions that you or other third-parties create in order to extend the functionality of Reporting Services. You must grant FullTrust to these extensions or assembly code in the policy configuration files associated with the Reporting Services component you're extending. Extensions shipped as a part of Reporting Services are signed with the report server public key and receive the FullTrust permission set.


You must modify the Reporting Services policy configuration files to allow FullTrust for any third-party extensions. If you do not add a code group with FullTrust for your custom extensions, they cannot be used by the report server.

For more information about the policy configuration files in Reporting Services, see Use Reporting Services security policy files.

Expressions Used in Reports

Report expressions are inline code expressions or user-defined methods contained within the Code element of a report definition language file. There's a code group that is already configured in the policy files that grants these expressions the Execution permission set by default. The code group looks like the following example:

   Description="This code group grants default permissions for code in report expressions and Code element. ">  

Execution permission allows code to run (execute), but not to use protected resources. All expressions found within a report are compiled into an assembly (called an "expression host" assembly) that is stored as a part of the compiled report. When the report is executed, the report server loads the expression host assembly and makes calls into that assembly to execute expressions. Expression host assemblies are signed with a specific key that is used to define the code group for all expression hosts.

Report expressions reference report object model collections (fields, parameters, etc.) and perform simple tasks like arithmetic and string operations. Code that performs these simple operations only requires Execution permission. By default, user-defined methods in the Code element and any custom assemblies are granted Execution permission in Reporting Services. Thus, for most expressions, the current configuration doesn't require that you modify any security policy files. To grant extra permissions to expression host assemblies, an administrator needs to modify the policy configuration files of the report server and Report Designer, and change the report expressions code group. Because it's a global setting, changing default permissions for the expression hosts affects all reports. For this reason, you should place all code that requires extra security into a custom assembly. Only this assembly is granted the permissions you need.


Code that calls external assemblies or protected resources should be incorporated into a custom assembly for use in reports. Doing so gives you more control over the permissions requested and asserted by your code. You should not make calls to secure methods within the Code element. Doing so requires you to grant FullTrust to the report expression host and grants all custom code full access to the CLR.


Do not grant FullTrust to the code group for a report expression host. If you do, you enable all report expressions to make protected system calls.

Custom assemblies referenced in reports

Some report expressions can call other code assemblies, also known in Reporting Services as custom assemblies. The report server expects these assemblies to have at least Execution permission in the policy configuration files. By default, policy files that ship with Reporting Services grant Execution permission to all assemblies starting from the 'My Computer' zone. You can grant extra permissions to custom assemblies as needed.

In some cases, you might need to perform an operation that requires specific code permissions in a report expression. Typically, this scenario means that a report expression needs to make a call to a secured CLR library method (such as one that accesses files or the system registry). The .NET Framework documentation describes the code permissions that are required to make this secure call. To execute the call, the calling code must be granted these specific, secure permissions. If you make the call from a report expression or the Code element, the expression host assembly must be granted the appropriate permissions. However, once you grant the expression host the permissions, all code that runs in any expression in any report is now granted that specific permission. It's much more secure to make the call from a custom assembly and grant that custom assembly the specific permissions.