
They can coexist but need a compatible path
IPv4 and IPv6 can coexist within the same infrastructure, but they are not directly interchangeable at the network layer. Communication requires the client and destination to share a usable family or rely on an appropriate intermediate mechanism. That is the practical answer to compatibility. Merely having an IP address on each side is insufficient: a coherent path must connect them. Architecture decisions should identify the family available at the client, transport, and service without assuming that every combination operates automatically or requires the same transition method.
Imagine an application querying two services: one accepts IPv6 and the other is accessible only over IPv4. The project does not need one answer for both. It can use IPv6 for the first and a compatible IPv4 path for the second, depending on available resources. Document the choice and observe the result. This avoids making a protocol difference into a mandatory simultaneous migration. It also clarifies that compatibility does not imply access authorization, guaranteed performance, or equivalent availability of every function offered by a destination.
198.51.100.10:8080203.0.113.24198.51.100.20:80802001:db8::24Dual stack serves destinations in both families
In a dual-stack deployment, the involved components support IPv4 and IPv6. A particular connection uses the family selected for its destination; it does not mix both formats inside one packet. This model can support coexistence with heterogeneous services, but requires configuring and operating both paths. DNS, routing, filtering, and monitoring must match the design.
Having both stacks on a device does not guarantee working external connectivity through each or establish that an application can use them correctly under normal and failure conditions. This describes the technical concept of dual stack, not the delivery of VLSM.io services. Each VLSM.io service provides one egress family, IPv4 or IPv6, contracted separately.
Automatic client selection must also be understood. Some clients try alternatives to reduce delay when one family encounters problems, while others depend on configuration or library-specific behavior. An application working on a workstation is therefore not automatically validated on its production server. Confirm the family used and test controlled failures. Operations should distinguish native IPv6 success from success achieved through IPv4 fallback so that a path still supporting important functionality is not removed prematurely based on an incomplete interpretation of application availability.
Translation serves defined scenarios
NAT64 is a translation mechanism used in scenarios where IPv6 clients need to reach IPv4 services. DNS64 can participate by supplying synthesized answers suitable for that path. Their functions differ: one translates traffic while the other operates in name resolution. The combination must be planned and validated in the environment. It should not be presented as universal conversion solving every direction, application, or addressing pattern. Source conditions, destination support, and dependencies determine whether this approach meets the requirements of a particular project.
Before choosing translation, investigate how the application obtains addresses and which protocols it uses. Literal-address configuration, independent resolution, or special dependencies may need additional handling. Assess capacity, session state, and observability of the intermediate component as well. Teams need to know where to investigate failures and which source the destination receives. Test the complete workflow rather than one simple request. An architecture may support ordinary web access while still requiring adjustments for a particular integration with different networking assumptions or application requirements.
198.51.100.10:8080203.0.113.24198.51.100.20:80802001:db8::24A proxy can separate inbound and outbound connections
A proxy can separate the two connections. At VLSM.io, IPv4 Proxy accepts the application at IPv4:Port and uses IPv4 egress; IPv6 Proxy also accepts the application at IPv4:Port but uses IPv6 egress. Support depends on the product, protocol, and configuration.
The application does not necessarily need direct connectivity to every destination using the same family as its proxy connection. However, the intermediary becomes part of the critical path. Authentication, limits, availability, and security must be defined because the solution does not eliminate dependencies; it concentrates them in a component that must be operated and observed correctly.
Distinguish a forward proxy used by clients to reach destinations from a reverse proxy placed in front of a published service. A dual-family reverse proxy can provide a compatible entry point while communicating with its backend through a family defined internally. This requires control over service publication and does not automatically modify third-party systems. Describe who administers the intermediary, where each connection terminates, and which properties must be preserved. The design should be verifiable through testing and authorized traffic records rather than inferred solely from product terminology.
Tunnels carry one protocol inside another
A tunnel encapsulates traffic to cross a network segment. Depending on the technology, this can carry one family over infrastructure using another externally. Encapsulation is not translation, however. After leaving the tunnel, the inner packet still needs a destination and network capable of handling it. This distinction is essential when comparing proposals: delivering a tunnel does not, by itself, make an IPv4-only server accept native IPv6 connections or resolve every application incompatibility. The endpoint beyond the tunnel still matters to the complete communication path.
Tunnel planning should include endpoints, routing, authentication where applicable, encapsulation overhead, and MTU. Test representative transfers and failure conditions, particularly when other intermediaries are present. A tunnel that establishes successfully may still carry only some expected flows because of routes or filters. Document the precise role of the delivery in the project. This lets the team distinguish failures in encapsulated transport, internal networking, and application behavior instead of treating every problem as a generic incompatibility between IPv4 and IPv6 that cannot be investigated further.
Choose based on the client and destination combination
If client and destination have working IPv6, a native connection may be sufficient. If both have working IPv4, that family also offers a direct path. When no common family exists, investigate translation, proxying, or changes to service publication according to available control.
Do not choose technology solely by address quantity or supplier preference. Consider application protocols, security, maintenance, and failure behavior. A straightforward, observable design is generally easier to sustain than a chain of undocumented intermediaries whose combined behavior is difficult to reproduce or support.
A decision matrix should record the current situation and intended outcome. Include who can change each component and which dependencies are external. An IPv4-only client accessing an IPv6-only service is not the same scenario as an IPv6 client reaching an IPv4 service through NAT64. Avoid reversing a solution without verifying support. Projects with many integrations may use different architectures for destination groups. Maintain common criteria, functional evidence, and documentation that allows choices to be reviewed as infrastructure and service capabilities evolve over time.
| Scenario | Path to evaluate |
|---|---|
| Client and service support IPv6 | Native IPv6 with workflow validation |
| Client and service support IPv4 | Direct IPv4 through the available network |
| IPv6 client, IPv4 service | NAT64/DNS64 or a suitable proxy, depending on the application |
| IPv4 client, IPv6 service | Suitable intermediary or dual-stack publication; do not assume reversed NAT64 |
Validate the service, not merely the technology name
Acceptance should demonstrate resolution, connection, transport security, and completion of required functions. Observe the source presented to the destination and session continuity too. Where an intermediary exists, monitor capacity and the errors it produces. Test unavailability and confirm that attempts are bounded and understandable. Compatibility findings need to be reproducible: another person should be able to identify the path and repeat verification without relying on assumptions about which family a client automatically selected or what an intermediary silently changed during the test.
VLSM.io can support connectivity and delivery planning for this scenario, considering IPv4, IPv6, and IP Proxy according to stated requirements. Bring the client-destination combination, application protocol, and operational targets to the discussion. The right choice is not simply declaring the families compatible or incompatible. It is building a supported, documented, tested path for the application. With this perspective, protocol coexistence becomes an engineering decision that can evolve in a controlled manner while preserving continuity and operational clarity for the people maintaining the service.
Institutional references
Reference documents on the topics covered in this article.
- IETFRFC 4213: Basic Transition Mechanisms for IPv6 Hosts and Routersdatatracker.ietf.org
- IETFRFC 6146: Stateful NAT64: Network Address and Protocol Translation from IPv6 Clients to IPv4 Serversdatatracker.ietf.org
- IETFRFC 6147: DNS64: DNS Extensions for Network Address Translation from IPv6 Clients to IPv4 Serversdatatracker.ietf.org
- IETFRFC 8305: Happy Eyeballs Version 2: Better Connectivity Using Concurrencydatatracker.ietf.org

