What Is Teredo? Tunneling, Teredo IP Addresses and the Xbox Error
What Teredo is, how Teredo tunneling carries IPv6 over UDP through NAT, how to decode a Teredo IP address, and what the Xbox “unable to qualify” error means.
In this guide
Teredo is an IPv6 transition technology, defined in RFC 4380, that carries IPv6 packets inside UDP over IPv4 so a device behind a NAT can reach the IPv6 internet without native IPv6. Every Teredo address starts with 2001:0: and encodes the Teredo server’s IPv4 address, a flags field, and the client’s obfuscated public IPv4 address and UDP port.
Test it now
CheckingRunning against our own test servers. Nothing is stored.
Open the IPv6 testWhat is Teredo and why does it exist?
Teredo is a service that lets hosts behind IPv4 NATs “obtain IPv6 connectivity by tunneling packets over UDP,” as RFC 4380 puts it. Christian Huitema wrote the RFC, published in February 2006 on the Standards Track. Microsoft shipped a client in Windows, which is why most of these addresses you will ever see belong to Windows PCs.
The problem it solves is specific. Earlier mechanisms such as 6to4 put IPv6 packets directly inside IPv4 packets (IPv4 protocol number 41), which home NAT routers cannot translate. UDP, by contrast, is something every NAT handles. So the IPv6 packet goes inside UDP, the NAT creates a mapping for that flow, and a public server tells the client what its mapping looks like.
RFC 4380 is explicit that this is a stopgap. Section 3.2 calls it an “IPv6 access of last resort,” and section 3.2.1 says nodes “SHOULD only use the Teredo service as a ‘last resort’ option,” preferring native IPv6, a 6to4 router at the NAT, or a configured tunnel. The authors even “expect to see a point in time at which the Teredo technology ceases to be used.” If your ISP gives you native IPv6, the tunnel has no job to do.
How does Teredo tunneling work?
Teredo has three roles, all defined in RFC 4380 section 5.
- Client. Your device. It owns a “service port,” a UDP port on its (usually private) IPv4 address, and sends every tunneled packet from that port.
- Server. A stateless host listening on UDP port 3544 (RFC 4380 section 2.7). It answers the client’s start-up requests and tells it the public IPv4 address and port the NAT mapped for it. It is not meant to carry the bulk of the traffic.
- Relay. A router that advertises 2001:0000::/32 into the IPv6 internet and forwards packets between native IPv6 hosts and clients, adding or removing the UDP wrapper as needed.
The start-up exchange is called qualification (RFC 4380 section 5.2.1). The client sends an IPv6 Router Solicitation, wrapped in UDP, to the server. If a Router Advertisement comes back, the client reads the “origin indication” in it to learn its mapped public IPv4 address and port. It then repeats the exchange with the server’s secondary address. Same address and port both times means a cone or restricted NAT, and the client is qualified. If they differ, RFC 4380 says “the client detects a symmetric NAT: it cannot use the Teredo service.” This is the mechanism behind the Xbox error discussed below.
To keep the NAT mapping alive and open paths to peers, clients and relays send bubbles: IPv6 packets with no payload (RFC 4380 section 2.8). A client assumes its mapping lasts 30 seconds without traffic (section 2.15) and refreshes on about that interval.
What is a Teredo IP address, and how do you decode one?
A Teredo IP address is a global IPv6 address inside 2001:0000::/32 whose remaining 96 bits are a packed record rather than an ordinary interface identifier. RFC 4380 section 4 defines five fields, in order:
| Field | Bits | Content |
|---|---|---|
| Prefix | 32 | 2001:0000 |
| Server IPv4 | 32 | IPv4 address of the Teredo server |
| Flags | 16 | NAT and address type; cone bit is the top bit, other bits randomized since RFC 5991 |
| Port | 16 | Client’s mapped (public) UDP port XOR 0xFFFF |
| Client IPv4 | 32 | Client’s mapped (public) IPv4 address XOR 0xFFFFFFFF |
The XOR “obfuscation” exists because some NATs rewrite any IPv4 address they spot inside a payload (RFC 4380 section 3.1). Inverting every bit hides the address from such NATs; it hides nothing from a person.
Here is a worked example, built with the documentation address ranges from RFC 5737 so it points at nobody. Suppose a client whose public address is 198.51.100.7, mapped to UDP port 40000, qualifies against a server at 192.0.2.45 with the cone bit clear. Its address, fully expanded, is:
2001:0000:c000:022d:0000:63bf:39cc:9bf8
which compresses to 2001:0:c000:22d:0:63bf:39cc:9bf8. Reading it back:
2001:0000is the Teredo prefix.c000:022dis the server: 0xc0 = 192, 0x00 = 0, 0x02 = 2, 0x2d = 45, so 192.0.2.45.0000is the flags field. Bit 15 (the cone bit, “C” in RFC 4380) is 0.63bfis the obfuscated port. 0x63bf XOR 0xffff = 0x9c40 = 40000.39cc:9bf8is the obfuscated client address. 0x39cc9bf8 XOR 0xffffffff = 0xc6336407, and 0xc6 = 198, 0x33 = 51, 0x64 = 100, 0x07 = 7, so 198.51.100.7.
Two caveats when you decode real addresses. First, RFC 5991 changed the flags field: 12 of the 13 unused bits are now random “to guard against address-scanning risks,” so on a modern client they look like noise rather than 0x0000 or 0x8000. RFC 5991 also sets the cone bit to zero and tells peers to ignore it. Second, the decoded IPv4 address is the NAT’s public address, not the LAN address; a device on 192.168.1.20 behind a router at 203.0.113.9 encodes 203.0.113.9.
Teredo vs 6to4 vs native IPv6
Both tunnels embed an IPv4 address in the IPv6 address, but they target different situations.
6to4 (RFC 3056) uses the prefix 2002::/16 and places the site’s IPv4 address directly after it (2002:V4ADDR::/48). That address “MUST NOT be a private address,” so 6to4 only works where a public IPv4 address is held. Packets travel as IPv4 protocol 41 with no UDP wrapper: less overhead, but no way through a NAT. In May 2015, RFC 7526 deprecated the anycast relay mechanism (192.88.99.0/24, RFC 3068), concluding that 6to4 “is unsuitable for widespread deployment and use in the Internet when used in its anycast mode.” It recommended that “future products should not support 6to4 anycast.” The basic 2002::/16 mechanism was left formally undeprecated.
Teredo (RFC 4380) uses 2001:0000::/32, wraps IPv6 in UDP so it crosses NATs, and depends on a server during qualification and a relay for traffic to native IPv6 hosts. It fails on symmetric NAT.
Native IPv6 needs neither. Your ISP delegates a prefix to your router, every device gets a global address, and the path to an IPv6 site is a normal route. RFC 4380 and RFC 6169 both say native access should be preferred whenever it exists. The guide to IPv4 vs IPv6 (https://checkmyip.com/learn/ipv4-vs-ipv6) covers how to tell which you have.
Teredo on Windows: states and netsh commands
Windows manages the tunnel through the netsh interface teredo context and, in PowerShell, the NetworkTransition module. The Microsoft Learn netsh interface reference (applies to Windows 10, Windows 11 and Windows Server 2016 through 2025) documents these commands:
netsh interface teredo show statedisplays the current Teredo state.netsh interface teredo set state type=<value>changes it, where the type is one ofdisabled,client,enterpriseclient,natawareclient,serverordefault. Optionalservername=,refreshinterval=andclientport=arguments set the Teredo server, the refresh interval in seconds and the client UDP port.
Microsoft’s parameter descriptions are brief. The Windows Server 2008 R2 netsh reference puts them as: disabled “Disables the Teredo service,” client “Enables the Teredo client,” and enterpriseclient “Skip managed network detection.” That last one matters. Per the Set-NetTeredoConfiguration reference the default type behaves as a client, but Windows first checks whether it is on a managed (domain) network. RFC 6169 section 3.1.3 notes that Windows disables the tunnel by default when it detects an enterprise domain controller; enterpriseclient bypasses that check. The PowerShell equivalent is Set-NetTeredoConfiguration -Type Client (or Disabled). Run both netsh and PowerShell as an administrator.
Microsoft’s pages do not publish a default server hostname or refresh interval; show state on your own machine prints the values in use.
The Xbox “Teredo is unable to qualify” error
On Windows, the Xbox app’s networking page reports a NAT type for multiplayer and party chat. When it shows “Teredo is unable to qualify,” it is reporting a failure of the RFC 4380 qualification exchange described above: the client sent its Router Solicitations to the server and did not end up qualified.
That leaves a short list of causes:
- The client is disabled. If
netsh interface teredo show stateshows a type ofdisabled, nothing is sent at all. Set it back withnetsh interface teredo set state type=default(administrator command prompt), then re-check. - UDP to port 3544 is blocked outbound. A firewall, a corporate or school network policy, or some VPN clients block the flow. RFC 6169 section 3.1.2 notes that blocking “destination UDP port 3544” is exactly how administrators stop the tunnel on purpose. If you cannot change that policy, it will not qualify on that network.
- Managed-network detection. A domain-joined PC may leave the client off; the
enterpriseclienttype skips that detection, but on a work machine that decision belongs to the administrator. - Symmetric NAT. If the NAT allocates a different public port for each destination, RFC 4380 says the client “cannot use the Teredo service.” This is a router or ISP behavior, not a Windows setting; the NAT type section below shows how to test for it.
- Double NAT or CGNAT. Two NAT layers make cause 4 more likely and make port forwarding ineffective. The double NAT guide (https://checkmyip.com/learn/double-nat) explains how to detect and fix that.
Microsoft’s Xbox support site has a troubleshooting article for this message. The page is a JavaScript application and our fetch returned only its shell, so this guide does not quote its steps; it is linked in the sources for you to open in a browser. The list above comes from RFC 4380 and the Microsoft Learn command references instead.
What does a Teredo address mean when it shows up in the IPv6 test?
Run https://checkmyip.com/ipv6-test. The tool reports whether your connection has working IPv6, shows the IPv6 address the site saw, and tells you whether your IPv4 and IPv6 traffic arrived from the same network. If the IPv6 address begins with 2001:0: (that is, it falls in 2001:0000::/32), you are reaching the site through the tunnel rather than native IPv6.
Three things follow from that:
- Your ISP is not giving you usable IPv6. Windows fell back to a last-resort tunnel. The fix is at the router or ISP level, not in Windows.
- The IPv4 and IPv6 “networks” will differ. Your IPv4 arrives from your ISP, while your IPv6 arrives via a relay run by someone else, so the ASN behind each address is different. The tool’s same-network check will reflect that.
- The address leaks your public IPv4. Anyone who sees it can decode your NAT’s public IPv4 address and port with the steps above. That is one reason such an address can defeat tools that only mask IPv4; the IPv6 leak guide (https://checkmyip.com/learn/ipv6-leak) covers that scenario.
If the test shows a native global address outside 2001:0::/32, the tunnel is not in use and you can disable it with no loss.
Teredo and your NAT type
Qualification depends on how your NAT maps UDP ports, which is the same property game consoles summarize as Open, Moderate or Strict. RFC 4380 section 3.1, using the RFC 3489 NAT classes, distinguishes a “cone” NAT, which reuses one public port for a given internal port regardless of destination, from a “symmetric” NAT, which allocates a new public port per destination. The tunnel works through cone, restricted and port-restricted NATs and fails on symmetric NATs.
https://checkmyip.com/nat-type-test measures exactly this. It sends your browser’s traffic to two STUN ports and compares the public port each one sees. If both see the same port, your NAT is endpoint-independent (cone-like, the “open” or “moderate” family). If they differ, it is address- and port-dependent, the symmetric behavior the tunnel cannot qualify through. If it reports port-dependent mapping, no Windows setting will fix qualification; see the guide to changing your NAT type (https://checkmyip.com/learn/how-to-change-nat-type) for the router-side options, including UPnP and bridging a double NAT.
Is Teredo a security risk?
RFC 4380 answers this itself in section 7.1, titled “Opening a Hole in the NAT”: “The very purpose of the Teredo service is to make a machine reachable through IPv6. By definition, the machine using the service will give up whatever firewall service was available in the NAT box.” Services listening on the Teredo address “will become the potential target of attacks from the entire IPv6 Internet.” The RFC’s mitigations: link-local services stay unreachable, a local (host) firewall “is recommended,” and IPsec can protect traffic end to end.
RFC 6169, which surveys tunnel security, adds detail. Tunneled traffic “may not receive the intended level of inspection” from network devices unless they are tunnel-aware (section 2.1.1). Public tunnels “essentially disable the filtering behavior of the NAT for all tunnel client ports” (section 4.3.1), and it “recommends that an IPv6 host firewall be used on all Teredo clients” (section 4.1.2). It also notes that “seeing a Teredo address suggests that the host it is on is probably running Windows” (section 5.2.2).
The practical reading: the tunnel does not break your router’s IPv4 firewall rules, but it creates an IPv6 path those rules never see. With Windows Firewall enabled, that path is filtered on the host; if you rely on the router alone, the client is reachable from the whole IPv6 internet. If you do not need it, disable it; if you need it for Xbox services, keep the host firewall on.
Sources
- RFC 4380: Teredo: Tunneling IPv6 over UDP through Network Address Translations (NATs)
- RFC 5991: Teredo Security Updates
- RFC 3056: Connection of IPv6 Domains via IPv4 Clouds (6to4)
- RFC 7526: Deprecating the Anycast Prefix for 6to4 Relay Routers
- RFC 6169: Security Concerns with IP Tunneling
- RFC 5737: IPv4 Address Blocks Reserved for Documentation
- Microsoft Learn: netsh interface (command reference, incl. netsh interface teredo)
- Microsoft Learn: Set-NetTeredoConfiguration (NetworkTransition PowerShell module)
- Microsoft Learn: Netsh commands for Interface IPv6 (set teredo / show teredo, Windows Server 2008 R2 reference)
- Xbox Support: troubleshooting article for the Teredo is unable to qualify message (page requires JavaScript; article text not readable by our fetcher)
Questions
What is Teredo tunneling?
Teredo tunneling wraps each IPv6 packet inside a UDP datagram carried over IPv4, so it can cross a NAT like any other UDP traffic. A Teredo server (UDP port 3544) helps the client learn its public address and port, and Teredo relays forward packets between Teredo clients and the native IPv6 internet.
Is it recommended to disable Teredo?
If you have native IPv6, or do not use anything that depends on Teredo, disabling it removes an interface that RFC 4380 itself calls a last resort and that bypasses your NAT’s filtering. If you use Xbox multiplayer or party chat on Windows, leave it enabled; Microsoft’s own error message appears when it cannot qualify.
What is a Teredo IP address?
An IPv6 address in 2001:0000::/32 built by a Teredo client. After the 32-bit prefix come the Teredo server’s IPv4 address (32 bits), 16 flag bits, the client’s public UDP port XOR 0xFFFF (16 bits) and the client’s public IPv4 address XOR 0xFFFFFFFF (32 bits). Anyone can decode the public IPv4 and port from it.
What does “Teredo is unable to qualify” mean?
Qualification is the RFC 4380 start-up exchange in which the client sends Router Solicitations over UDP to the Teredo server and learns its NAT type. The message means that exchange failed: no reply came back (UDP 3544 blocked, server unreachable, Teredo disabled) or the client detected a symmetric NAT, which Teredo cannot use.
Back to your own address: see your public IP

