Planning and getting started on the Windows Defender Device Guard deployment process

Applies to

  • Windows 10
  • Windows Server 2016

This topic provides a roadmap for planning and getting started on the Windows Defender Device Guard deployment process, with links to topics that provide additional detail. Planning for Windows Defender Device Guard deployment involves looking at both the end-user and the IT pro impact of your choices. Use the following steps to guide you.

Planning

  1. Review requirements, especially hardware requirements for VBS. Review the virtualization-based security (VBS) features described in How Windows Defender Device Guard features help protect against threats. Then you can assess your end-user systems to see how many support the VBS features you are interested in, as described in Hardware, firmware, and software requirements for Windows Defender Device Guard.

  2. Group devices by degree of control needed. Group devices according to the table in Windows Defender Device Guard deployment in different scenarios: types of devices. Do most devices fit neatly into a few categories, or are they scattered across all categories? Are users allowed to install any application or must they choose from a list? Are users allowed to use their own peripheral devices?
    Deployment is simpler if everything is locked down in the same way, but meeting individual departments’ needs, and working with a wide variety of devices, may require a more complicated and flexible deployment.

  3. Review how much variety in software and hardware is needed by roles or departments. When several departments all use the same hardware and software, you might need to deploy only one Windows Defender Application Control (WDAC) policy for them. More variety across departments might mean you need to create and manage more WDAC policies. The following questions can help you clarify how many WDAC policies to create:

    • How standardized is the hardware?
      This can be relevant because of drivers. You could create a WDAC policy on hardware that uses a particular set of drivers, and if other drivers in your environment use the same signature, they would also be allowed to run. However, you might need to create several WDAC policies on different "reference" hardware, then merge the policies together, to ensure that the resulting policy recognizes all the drivers in your environment.

    • What software does each department or role need? Should they be able to install and run other departments’ software?
      If multiple departments are allowed to run the same list of software, you might be able to merge several WDAC policies to simplify management.

    • Are there departments or roles where unique, restricted software is used?
      If one department needs to run an application that no other department is allowed, it might require a separate WDAC policy. Similarly, if only one department must run an old version of an application (while other departments allow only the newer version), it might require a separate WDAC policy.

    • Is there already a list of accepted applications?
      A list of accepted applications can be used to help create a baseline WDAC policy.
      As of Windows 10, version 1703, it might also be useful to have a list of plug-ins, add-ins, or modules that you want to allow only in a specific app (such as a line-of-business app). Similarly, it might be useful to have a list of plug-ins, add-ins, or modules that you want to block in a specific app (such as a browser).

    • As part of a threat review process, have you reviewed systems for software that can load arbitrary DLLs or run code or scripts? In day-to-day operations, your organization’s security policy may allow certain applications, code, or scripts to run on your systems depending on their role and the context. However, if your security policy requires that you run only trusted applications, code, and scripts on your systems, you may decide to lock these systems down securely with Windows Defender Application Control policies. You can also fine-tune your control by using Windows Defender Application Control in combination with AppLocker, as described in Windows Defender Device Guard with AppLocker.

      Legitimate applications from trusted vendors provide valid functionality. However, an attacker could also potentially use that same functionality to run malicious executable code that could bypass WDAC.

      For operational scenarios that require elevated security, certain applications with known Code Integrity bypasses may represent a security risk if you whitelist them in your WDAC policies. Other applications where older versions of the application had vulnerabilities also represent a risk. Therefore, you may want to deny or block such applications from your WDAC policies. For applications with vulnerabilities, once the vulnerabilities are fixed you can create a rule that only allows the fixed or newer versions of that application. The decision to allow or block applications depends on the context and on how the reference system is being used.

      Security professionals collaborate with Microsoft continuously to help protect customers. With the help of their valuable reports, Microsoft has identified a list of known applications that an attacker could potentially use to bypass Windows Defender Application Control. Depending on the context, you may want to block these applications. To view this list of applications and for use case examples, such as disabling msbuild.exe, see Deploy Windows Defender Application Control: steps.

  4. Identify LOB applications that are currently unsigned. Although requiring signed code (through WDAC) protects against many threats, your organization might use unsigned LOB applications, for which the process of signing might be difficult. You might also have applications that are signed, but you want to add a secondary signature to them. If so, identify these applications, because you will need to create a catalog file for them. For a basic description of catalog files, see the table in Introduction to Windows Defender Device Guard: virtualization-based security and Windows Defender Application Control. For more background information about catalog files, see Reviewing your applications: application signing and catalog files.

Getting started on the deployment process

  1. Optionally, create a signing certificate for Windows Defender Application Control. As you deploy WDAC, you might need to sign catalog files or WDAC policies internally. To do this, you will either need a publicly issued code signing certificate (that you purchase) or an internal CA. If you choose to use an internal CA, you will need to create a code signing certificate. For more information, see Optional: Create a code signing certificate for Windows Defender Application Control.

  2. Create WDAC policies from “golden” computers. When you have identified departments or roles that use distinctive or partly-distinctive sets of hardware and software, you can set up “golden” computers containing that software and hardware. In this respect, creating and managing WDAC policies to align with the needs of roles or departments can be similar to managing corporate images. From each “golden” computer, you can create a WDAC policy, and decide how to manage that policy. You can merge WDAC policies to create a broader policy or a master policy, or you can manage and deploy each policy individually. For more information, see:

  3. Audit the WDAC policy and capture information about applications that are outside the policy. We recommend that you use “audit mode” to carefully test each WDAC policy before you enforce it. With audit mode, no application is blocked—the policy just logs an event whenever an application outside the policy is started. Later, you can expand the policy to allow these applications, as needed. For more information, see Audit Windows Defender Application Control policies.

  4. Create a “catalog file” for unsigned LOB applications. Use the Package Inspector tool to create and sign a catalog file for your unsigned LOB applications. For more information, review step 4 Identify LOB applications that are currently unsigned, earlier in this list, and see Deploy catalog files to support Windows Defender Application Control. In later steps, you can merge the catalog file's signature into your WDAC policy, so that applications in the catalog will be allowed by the policy.

  5. Capture needed policy information from the event log, and merge information into the existing policy as needed. After a WDAC policy has been running for a time in audit mode, the event log will contain information about applications that are outside the policy. To expand the policy so that it allows for these applications, use Windows PowerShell commands to capture the needed policy information from the event log, and then merge that information into the existing policy. You can merge WDAC policies from other sources also, for flexibility in how you create your final WDAC policies. For more information, see:

  6. Deploy WDAC policies and catalog files. After you confirm that you have completed all the preceding steps, you can begin deploying catalog files and taking WDAC policies out of auditing mode. We strongly recommend that you begin this process with a test group of users. This provides a final quality-control validation before you deploy the catalog files and WDAC policies more broadly. For more information, see:

  7. Enable desired virtualization-based security (VBS) features. Hardware-based security features—also called virtualization-based security (VBS) features—strengthen the protections offered by Windows Defender Application Control, as described in How Windows Defender Device Guard features help protect against threats.

    Warning

    Virtualization-based protection of code integrity may be incompatible with some devices and applications. We strongly recommend testing this configuration in your lab before enabling virtualization-based protection of code integrity on production systems. Failure to do so may result in unexpected failures up to and including data loss or a blue screen error (also called a stop error).

    For information about enabling VBS features, see Enable virtualization-based protection of code integrity.