Security in Windows Server 2019

At the end of last December, Microsoft published a document titled What’s new in Windows Server 2019, covering the new features and renewed functionalities provided by the new version of Windows Server. This post will focus on those features related to security improvements provided by Windows Defender ATP that had already been seen in Windows 10 through Windows Defender Exploit Guard, EMET (Enhanced Mitigation Experience Toolkit, which ceased to have support last July 31, 2018), as well as WDAC (Windows Defender Application Control).

While writing this post, the research deepened and led to a much more comprehensive document on ATP, specifically Windows Defender Advanced Threat Protection. This post is intended to be a brief, orderly summary of the links in the aforementioned document.

Windows Defender Advanced Threat Protection (ATP)

The Windows Defender ATP system is designed to protect the kernel and system memory from malicious files and processes, either by blocking or terminating them, in order to prevent host intrusion. In its development, it relies on several aspects to be taken into account to reduce intrusion.

1. Attack Surface Reduction (ASR)

Attack Surface Reduction (ASR) is based on a complex set of rules managed by system administrators that allow to block potentially dangerous files based on their behavior. These rules block these files based mainly on the following aspects or behaviors:

  • Executable in mail.
  • Secondary processes, Win32 API calls from macros or executables created by Office applications, as well as the possibility of code injection.
  • Scripts that execute code downloaded by them, unsigned and unreliable processes from USB, …

2. Network Protection

This network filtering capability is located in the system kernel and is aimed at protecting the host. In this case, it blocks outgoing computer connections against potentially dangerous domains to prevent phishing damage, sites that may attempt to install malware on the machine, and even spread to other machines on the network. The blocking relies mainly on reputation intelligence based on both the IP and the domain name, combining online searches and caching. If the result indicates that the connection is to a bad reputation site, the Internet connection (to this web-based malware) is blocked, whether the call is generated from a browser or from a background process.

3. Controlled Access Folder

In recent years, ransomware attacks that led to the encryption of files in organizations and businesses have been notorious. To avoid these behaviors, controlled access to folders is available. In essence, the aim is to avoid the use of files by processes not defined as trusted. The management of this feature in the IT infrastructure can be carried out through GPO or PowerShell. In addition to controls definition, each time an attempt to make a change to protected directories is blocked, an alert is generated in Windows Defender ATP.

4. Exploit Protection

Vulnerability protection was already present in EMET and is now available in Windows Defender ATP. Since EMET is no longer supported, Microsoft has implemented mechanisms for migrating the rules defined in EMET to ATP. Among the mitigations it carries out are those related to code execution, remote image integrity validation, blocking of non-trusted sources, API validations, blocking of secondary processes and driver validation. As mentioned above, some were already present in EMET and others are their own or have been improved by ATP.

With respect to the four points cited so far, all these mechanisms can be enabled in audit mode, in which the system simulates its execution, leaving traces of the simulated actions in the event logs, so that after a given analysis period, it can be determined if the rules designed are valid and can go live. In addition, the administrator can enable the notification so that the user is aware of the locks. This is useful in case a false positive occurs, so the user can notify the error caused by the locking mechanism.

5. Windows Defender Application Control (WDAC)

As indicated in the article, WDAC appeared with Windows Server 2016. As a management improvement, Microsoft has developed Code Integrity (CI) policies so executables that put integrity at risk can be blocked. In this case, instead of considering all the applications and libraries as trustworthy, the approach is reversed and non-trust is assumed, so which applications users can run and which code can be executed in the system kernel must be established explicitly. On the other hand, it also incorporates the ability to block unsigned scripts and .msi, as well as determine whether add-ons or application modules can be executed, making use of simple rules that relate executables to libraries.

6. Hardware-based isolation

  • Application isolation

This model is based on the definition of the sites that the organization considers trustworthy. Therefore, access to sites not included will be considered non-trustworthy. When accessing a non-trusted site, access is done in an isolated container, so if the site is really malicious, the host will be protected against intrusion, since the container is anonymous and has no access to user credentials.

  • System isolation

Protect and maintain system integrity at startup

The system is prepared to prevent any bootkit from booting before the operating system boot loader. This is done using Root of Trust (RoT) based hardware, which is an element of the Unified Extensible Firmware Interface (UEFI). After this validation, Windows and firmware can be started.

Protect and maintain the integrity of the system after booting

Despite the mechanisms developed to avoid increasing privileges attacks, it is not yet certain that the integrity of the critical services of the operating system can be maintained. To do this Windows 10 introduced the security model based on virtualization (VBS). With this concept, it is possible to isolate data in a hardware-based model. In this way, critical services such as Credential Guard, Device Guard, Virtual TPM and parts of Windows Defender, etc., can be protected at runtime.

Validation of the integrity of the system locally and remotely

Through Trusted Platform Module 2.0 (TPM 2.0), the system obtains data that will serve as integrity measures. Both the process and the information are isolated from the hardware, so data collection is not subject to alteration. Through remote systems such as Intune or System Center Configuration Manager (SCCM) can be requested for analysis. If the analysis shows that the system is compromised, you could, for example, deny access to the resources offered by the compromised device.

As commented, what has been tried with this post is to make a first approach to Windows Defender ATP, offering an entry point of the options available to the system administrators, in order to maximize the levels of security offered by Windows Server 2019.

Bibliography and Resources