How to Identify Applications That Use Windows CLAT (WinCLAT) with Sysmon

Windows CLAT (WinCLAT) allows applications that use IPv4 sockets or literal IPv4 addresses to continue working on IPv6-only and IPv6-mostly networks by translating their IPv4 traffic to IPv6 for delivery to a NAT64 gateway. For administrators planning an IPv6-only deployment, this raises an obvious question: which applications still depend on IPv4 and are using WinCLAT?

Application Visibility

WinCLAT’s built-in logging does not identify the applications using it. This post explains how to use Microsoft Sysinternals Sysmon to record the executable, user account, and destination associated with IPv4 connections that use the WinCLAT source address. It also describes the filtering requirements and limitations of this approach. The behavior described applies to Sysmon 15.22. Later versions may differ.

What WinCLAT Logs

WinCLAT records events in the Microsoft-Windows-IPxlatCfg/Operational event log, but those events describe the WinCLAT instance on each interface, not the traffic passing through it.

Event IDDescription
1106Instance enabled
1107Instance disabled
1108Remote prefix discovered
1109Translation started
1110Translation stopped
1111WinCLAT IPv6 address DAD failure

No event identifies the application that used WinCLAT or the destination it connected to. The netsh.exe interface clat commands report the same kind of instance and configuration state. As a result, WinCLAT can translate traffic on an interface with no built-in record of what generated it.

Logging WinCLAT Connections with Sysmon

Sysmon is a Microsoft Sysinternals tool that runs as a Windows service and logs detailed system activity, including process creation, network connections, and file changes, to the Windows Event Log. Sysmon can be configured specifically to log network traffic generated by WinCLAT.

Identifying WinCLAT Traffic in Event ID 3

Sysmon Event ID 3 (Network connection) logs TCP and UDP connections and links each one to a process. When an application uses WinCLAT, its socket is still an IPv4 socket, so the event shows the application’s IPv4 addresses rather than the translated IPv6 addresses:

  • Calling executable and account – Image, ProcessId, ProcessGuid, and User
  • Destination – DestinationIp, DestinationPort, and Protocol
  • WinCLAT source address – SourceIp

The source address is what makes this useful. While WinCLAT is active, it assigns the interface an IPv4 address from the IPv4 Service Continuity Prefix 192.0.0.0/29 (RFC 7335), typically 192.0.0.1. Filtering on that source separates traffic that uses WinCLAT from other network activity. Traffic that uses a native IPv4 route has a different source address, and traffic from applications that use IPv6 directly, including through DNS64 and NAT64, is not IPv4 traffic and does not match.

Preparing the WinCLAT Filter

Network connection events are disabled by default, so Sysmon requires a configuration file to enable them. Save the following as sysmon-winclat.xml, or download it from GitHub here.

<Sysmon schemaversion="4.82">
  <DnsLookup>False</DnsLookup>
  <EventFiltering>
    <NetworkConnect onmatch="include">
      <SourceIp condition="is any">192.0.0.1;192.0.0.2;192.0.0.3;192.0.0.4;192.0.0.5;192.0.0.6;192.0.0.7</SourceIp>
    </NetworkConnect>
    <NetworkConnect onmatch="exclude">
      <DestinationIp condition="begin with">224.</DestinationIp>
      <DestinationIp condition="begin with">239.</DestinationIp>
      <DestinationIp condition="is">255.255.255.255</DestinationIp>
      <DestinationIp condition="is any">192.0.0.0;192.0.0.1;192.0.0.2;192.0.0.3;192.0.0.4;192.0.0.5;192.0.0.6;192.0.0.7</DestinationIp>
    </NetworkConnect>
    <ProcessCreate onmatch="include" />
    <FileCreateTime onmatch="include" />
    <ProcessTerminate onmatch="include" />
    <DriverLoad onmatch="include" />
    <RawAccessRead onmatch="include" />
  </EventFiltering>
</Sysmon>

The configuration does the following.

  • Includes connections with a source address in the WinCLAT range. A computer with several translating interfaces uses a different address on each one.
  • Excludes multicast and broadcast destinations, and destinations in the WinCLAT range itself. Local protocols such as LLMNR and NetBIOS can use the WinCLAT IPv4 address, but this traffic is not destined for NAT64. A source address in the WinCLAT range therefore does not by itself prove that a packet was translated.
  • Disables reverse DNS lookups with DnsLookup. A reverse lookup would add unnecessary queries, and its result would not show the name the application requested anyway.
  • Turns off other default event types with empty include rules. Without these rules, Sysmon also logs events such as process creation and termination, which are unrelated to WinCLAT and can far outnumber the network connection events. Sysmon still logs its service state and configuration change events, which cannot be filtered.

Turning off process creation events removes the command line and parent process from the log. To keep that context for investigations, remove the ProcessCreate rule and accept the additional volume. See the Sysmon documentation for more filtering options.

Installing and Configuring Sysmon

Sysmon is available natively in Windows 11 since February, 2026, but it is not installed by default. Administrators can install it by navigating to Settings > System > Optional features > More Windows features and selecting Sysmon.

Optionally, administrators can install Sysmon by opening an elevated PowerShell command window and running the following command.

Enable-WindowsOptionalFeature -Online -FeatureName Sysmon

For Windows versions that don’t include Sysmon, administrators can download Sysmon from the Sysinternals website.

Applying the WinCLAT Filter

After installing Sysmon, open an elevated command window and install it with the configuration file by running the following command.

sysmon.exe -i sysmon-winclat.xml

To apply a changed configuration to a computer where Sysmon is already installed, use -c.

sysmon.exe -c sysmon-winclat.xml

Reading Sysmon Events

Events are recorded in the Microsoft-Windows-Sysmon/Operational log.

Sysmon event log indicating the calling application.

Sysmon event log indicating the source and destination IPv4 addresses used by the calling application.

Viewing Recent Events with PowerShell

Run the following command to display the time and message for up to 20 of the most recent Sysmon Event ID 3 records.

Get-WinEvent -FilterHashtable @{ LogName = 'Microsoft-Windows-Sysmon/Operational'; Id = 3 } -MaxEvents 20 | Format-List -Property TimeCreated, Message

Sysmon event log details viewed using PowerShell.

The example above shows several fields of interest for a TCP connection to an IPv4 address through WinCLAT.

Network connection detected:
ProcessId: 10536
Image: C:\Tools\putty.exe
User: LAB\rhicks
Protocol: tcp
Initiated: true
SourceIsIpv6: false
SourceIp: 192.0.0.1
SourcePort: 50935
DestinationIsIpv6: false
DestinationIp: 172.16.0.254
DestinationPort: 22

Listing Applications and User Accounts

To produce a list of the applications and destinations recorded on a computer, read the values from the event data and remove the duplicates.

Get-WinEvent -FilterHashtable @{ LogName = 'Microsoft-Windows-Sysmon/Operational'; Id = 3 } |
    Select-Object -Property @{ Name = 'Image'; Expression = { $_.Properties[4].Value } },
        @{ Name = 'User'; Expression = { $_.Properties[5].Value } } |
    Sort-Object -Property Image, User -Unique |
    Format-Table -AutoSize

This query includes all retained Event ID 3 records, including events collected before the WinCLAT configuration was applied.

Unique application and user account combinations recorded by Sysmon.

To report across many computers, collect the events centrally with Windows Event Forwarding or a SIEM agent.

Interpreting the Results

Sysmon Event ID 3 records different information depending on the protocol and whether the connection succeeds. The table below shows which types of traffic can produce an event and what each record tells you about the connection.

Traffic through WinCLATRecorded by Event ID 3What an event means
TCP connectionYesThe connection was established
TCP attempt with no responseMay not be recordedNot applicable
UDPYes, with or without a replyThe application attempted to send, not that the traffic was delivered
ICMPNoNot applicable

Limitations of Event ID 3

Use these events to identify which processes communicate with IPv4 destinations. When reviewing the results, keep in mind that some activity may be missing and that the recorded fields provide limited information about the application’s behavior.

  • ICMP is not recorded. Event ID 3 covers TCP and UDP only, so ping to an IPv4 address leaves no event.
  • Failed TCP attempts can be missed. An application that repeatedly tries to reach an unresponsive IPv4 address may not appear in the log. This matters because unreachable destinations through WinCLAT time out rather than failing immediately.
  • Events show WinCLAT use, not literal use. An application that resolves a name to an IPv4 address and connects with an IPv4 socket produces the same event as an application that uses a literal IPv4 address. The destination is the resolved address in both cases, so Sysmon cannot confirm that an address is hard-coded. Sysmon Event ID 22 (DNS query) can provide supporting context, because it identifies the process that made a query, but the absence of a matching query is not proof of a literal. The configuration above does not log this event, so collecting it requires adding a DnsQuery rule.
  • Event counts are not traffic counts. Sysmon does not record one event per packet or per flow, so several UDP sends on one socket can produce a different number of events. Use the events to identify which applications communicate with which destinations, not to measure traffic volume.
  • Shared processes hide the service. Traffic from a Windows service can appear with an Image of svchost.exe. Use the ProcessId to identify the service while the process is still running.
  • Native IPv4 is excluded. On a computer with native IPv4 connectivity, some connections may use a native IPv4 source address instead of the WinCLAT source address. Those connections do not match this filter and are not included in the results.

Additional Troubleshooting Tools

If traffic that Sysmon does not record becomes important, such as ICMP or failed connection attempts, the following sources may help. Each has its own trade-offs, so evaluate them in your environment.

  • Windows Filtering Platform auditing. The Filtering Platform Connection audit subcategory records permitted and blocked connections in the Security event log as events 5156 and 5157, including the application where available. It can generate a high volume of events.
  • Packet capture. A capture on the physical adapter with a tool such as pktmon.exe can show translated IPv6 traffic, including ICMP and unanswered connection attempts. The IPv4 destination is embedded in the destination address under the NAT64 prefix. Packet headers alone do not identify the calling executable, so process attribution requires additional correlation.
  • NAT64 gateway logs. Depending on the gateway, its logs can show the client and destination of translated traffic, but not the application.

Summary

WinCLAT logs instance state changes but does not identify the applications using it. Sysmon Event ID 3, filtered on WinCLAT source addresses and with common local traffic excluded, provides a practical way to identify applications likely using WinCLAT. The events record the executable, user account, destination address, and port for matching TCP and UDP connections. They do not cover ICMP, may omit failed TCP attempts, and cannot distinguish a literal IPv4 address from an address obtained through name resolution. Use the results to identify applications for further investigation and confirm translation with a packet capture when needed.

Additional Information

Sysmon XML Configuration File for WinCLAT Tracing on GitHub

WinCLAT resources on directaccess.richardhicks.com

Always On VPN NPS Auditing and Logging

The Network Policy Server (NPS) event log is incredibly valuable for administrators when troubleshooting Always On VPN user tunnel connectivity issues. Administrators can find these pertinent events by opening the Event Viewer on the NPS server (eventvwr.msc) and navigating to Custom Views > Server Roles > Network Policy and Access Services.

Event Logs

When configured correctly, event logs will record the disposition of all authentication requests, allowed or denied. The two most common recorded events are event IDs 6272 (access granted) and 6273 (access denied).

NPS Event ID 6272 – Access granted.

NPS Event ID 6273 – Access denied.

Auditing

In some cases, administrators may find none of these events recorded even though user authentication is working correctly. Here, the only events recorded are NPS informational events indicating which domain controller the NPS server is using to perform authentication.

The lack of 6272 and 6273 events in the event log indicates that auditing for NPS events is not enabled. Open an elevated PowerShell window and run the following command to view the current auditing setting for NPS events.

auditpol.exe /get /subcategory:”Network Policy Server”

Open an elevated PowerShell window and run the following command to enable auditing for NPS events.

auditpol.exe /set /subcategory:”Network Policy Server” /success:enable /failure:enable

Group Policy

Alternatively, consider using Active Directory group policy to enforce the NPS server auditing settings. Open the Group Policy Management Console (GPMC) and create a new GPO. Next, navigate to Computer Configuration > Policies > Windows Settings > Security Settings > Advanced Audit Policy Configuration > Audit Policies > Logon/Logoff > Audit Network Policy Server and select the option to audit both success and failure attempts.

Once complete, link this GPO to the OU where the NPS servers reside.

Missing Events

If auditing is enabled and there are no recorded 6272 or 6273 events, the NPS server did not receive any authentication requests from the VPN server. Review the event logs on any other NPS servers if there is more than one configured. In addition, this may indicate that network communication between the VPN and NPS server is blocked. Ensure network connectivity and name resolution are working as expected.

Troubleshooting Guides

Are you interested in learning more about Always On VPN troubleshooting? My Always On VPN book contains an entire chapter dedicated to troubleshooting. Also, my Always On VPN video training course on Pluralsight includes a module on troubleshooting. The video training course is available to Pluralsight subscribers only. If you don’t have a Pluralsight subscription, you can sign up for a free trial here.

Additional Information

Troubleshooting Always On VPN Errors 691 and 812

Troubleshooting Always On VPN Errors 691 and 812 – Part 2

Troubleshooting Always On VPN Errors 691 and 812 – Part 3

Always On VPN NPS Load Balancing

Always On VPN Split vs. Force Tunneling

Always On VPN Split vs. Force TunnelingDuring the planning phase of a Windows 10 Always On VPN implementation the administrator must decide between two tunneling options for VPN client traffic – split tunneling or force tunneling. When split tunneling is configured, only traffic for the on-premises network is routed over the VPN tunnel. Everything else is sent directly to the Internet. With force tunneling, all client traffic, including Internet traffic, is routed over the VPN tunnel. There’s been much discussion recently on this topic, and this article serves to outline the advantages and disadvantages for both tunneling methods.

Force Tunneling

Force tunneling is typically enabled to meet the following requirements.

Visibility and Control

By routing all the client’s Internet traffic over the VPN tunnel, administrators can inspect, filter, and log Internet traffic using existing on-premises security solutions such as web proxies, content filters, or Next Generation Firewalls (NGFW).

Privacy

Enabling force tunneling ensures privacy and protection of all Internet communication. By routing all Internet traffic over the VPN, administrators can be certain that all communication from the Always On VPN client is encrypted, even when clients access unencrypted web sites or use untrusted or insecure wireless networks.

Force Tunneling Drawbacks

While configuring force tunneling for Always On VPN has some advantages, it comes with some serious limitations as well.

Poor User Experience

User experience is often degraded when all Internet traffic is routed over the VPN. These suboptimal network paths increase latency, and VPN encapsulation and encryption overhead increase fragmentation, leading to reduced throughput. Most Internet traffic is already encrypted in some form, and encrypting traffic that is already encrypted makes the problem even worse. In addition, force tunneling short-circuits geographic-based Content Delivery Networks (CDNs) further reducing Internet performance. Further, location-based services are often broken which can lead to improper default language selection or inaccurate web search results.

Increased Resource Consumption

Additional resources may need to be provisioned to support force tunneling. With corporate and Internet traffic coming over the VPN, more CPU, memory, and network resources may be required. Deploying additional VPN servers and higher throughput load balancers to support the increase in network traffic may also be necessary. Force tunneling also places higher demands on Internet Service Provider (ISP) links to the corporate datacenter.

Split Tunneling

The alternative to force tunneling is “split tunneling”. With split tunneling configured, only traffic destined for the internal corporate network is routed over the VPN. All other traffic is sent directly to the Internet. Administrators define IP networks that should be routed over the VPN, and those networks are added to the routing table on the VPN client.

Security Enforcement

The challenge of providing visibility and control of Internet traffic with split tunneling enabled can be met using a variety of third-party security solutions. Microsoft Defender ATP recently introduced support for web content filtering. Also, there are numerous cloud-based security offerings from many vendors that allow administrators to monitor and control client-based Internet traffic. Zscaler and Cisco Umbrella are two popular solutions, and no doubt there are many more to choose from.

Recommendations

The general guidance I provide customers is to use split tunneling whenever possible, as it provides the best user experience and reduces demands on existing on-premises infrastructure. Enabling split or force tunneling is ultimately a design decision that must be made during the planning phase of an Always On VPN implementation project. Both configurations are supported, and they each have their merits.

In today’s world, with many applications accessible via public interfaces, force tunneling is an antiquated method for providing visibility and control for managed devices in the field. If required, investigate the use of Microsoft or other third-party solutions that enforce security policy in place without the requirement to backhaul client Internet traffic to the datacenter over VPN for inspection, logging, and filtering.

Additional Information

Whitepaper: Enhancing VPN Performance at Microsoft

Whitepaper: How Microsoft Is Keeping Its Remote Workforce Connected

Microsoft Defender ATP Web Content Filtering

DirectAccess SQL Server High CPU Usage

UPDATE – March 14, 2016: Microsoft has published official guidance for implementing the changes outlined in this article using PowerShell. Details here.

Introduction

DirectAccess SQL Server High CPU UsageRADIUS and Inbox accounting are the two supported logging options for DirectAccess in Windows Server 2012 R2. When Inbox accounting is selected, a Windows Internal Database (WID) is provisioned. Part of the base operating system, WID is functionally similar to SQL Server Express.

SQL Server Utilization Issues

Over the last few months I’ve had a few customers reach out to me with a peculiar performance issue. For customers with very busy DirectAccess servers, where those servers have also been configured to use Inbox accounting, they’ve reported observing unusually high CPU utilization on the sqlservr.exe process.

DirectAccess SQL Server High CPU Usage
Image courtesy Thomas Vuylsteke. Used with permission. – setspn.blogspot.com

As luck would have it, Thomas Vuylsteke, a Microsoft Platforms Premiere Field Engineer (PFE), had already identified the issue and a workaround. Thomas traced the source of high CPU utilization on the sqlservr.exe process to a missing index on a session state table in the DirectAccess accounting database. If you are interested in learning how he performed the troubleshooting to identify and resolve this problem, you can read his entire blog post here.

Resolution

To resolve this issue, create an index on the Session Table in the DirectAccess database. Changes to WID must be made locally, as it is not remotely manageable. WID does not include a management interface, which means the SQL Server management tools would normally have to be installed. However, I’m not a fan of installing any extraneous software on the DirectAccess server, so thankfully one of the readers of Thomas’ excellent article on this subject, Fredrik Elmqvist, provided a very helpful alternative. Fredrik suggesting using the HeidiSQL tool, for which a fully portable version exists. This allows for changes to be made to the WID database without having to install any additional software.

Changes to WID

Begin by downloading the portable version of HeidiSQL here. Next, log on to the DirectAccess server as the local administrator. It is crucial that you must be the local administrator, not just a local or domain user with local administrator privileges. Extract the files from the download and copy them to the DirectAccess server, then follow these steps:

  1. Double-click heideisql.exe to launch the management tool.
  2. Click on New and then for the Network Type select Microsoft SQL Server (named pipe).
  3. For the Hostname / IP: enter \\.\pipe\MICROSOFT##WID\tsql\query.
  4. Select the option to Use Windows Authentication.
  5. Click Open to continue.DirectAccess SQL Server High CPU Usage
  6. Click the Query tab in the center console window and enter the following commands:
    Use RaAcctDb
    Create NonClustered Index IdxSessionTblSessionState on SessionTable (SessionState,ConnectionID)
  7. Click the Run icon in the tool bar or press F9. This will execute the code and create the missing index on the Session Table in the DirectAccess database.DirectAccess SQL Server High CPU Usage
  8. Confirm the index was created by clearing the previous query or creating a new query and then entering the following commands:
    select * from sys.indexes
    where name like ‘idx%’
    order by name ascDirectAccess SQL Server High CPU Usage

Summary
Once the change has been made, sqlservr.exe CPU utilization should return to normal. If you have multiple DirectAccess servers configured in a load-balanced array or in a multisite configuration, be sure to repeat these steps on each DirectAccess server in the organization.

Monitoring DirectAccess Machine and User Activity with Windows Component Event Logging

Monitoring DirectAccess Machine and User Activity with Component Event LoggingThe monitoring of DirectAccess machine and user activity presents some unique challenges for security administrators. All DirectAccess client communication destined for the internal corporate network is translated by the DirectAccess server and appears to originate from the DirectAccess server’s internal IPv4 address. Also, the public IPv4 address for DirectAccess clients using the IP-HTTPS IPv6 transition protocol is not visible using the native reporting tools. In addition, vital information such as source ports used by the DirectAccess server for internal connections and source ports used by DirectAccess clients is not available. This lack of granular connection logging creates a serious blind spot for administrators conducting forensic investigations.

As veteran Microsoft Premier Field Engineer (PFE) Martin Solis described in detail in a recent blog post, all of these details are in fact logged. However, gathering this information is not exactly intuitive. To collect this essential information it will be necessary to leverage Windows component event logging. By searching the IPHLPSVC, Base Filtering Engine Connections, Base Filtering Resource Flows, and WinNAT operational logs, it is possible to gather all of the information necessary for uniquely identifying DirectAccess corporate network communication.

Be sure to read Martin’s excellent article about using Windows component event logging to monitor DirectAccess machine and user activity, which can be found here.