Networks
IPv4 and IPv6: two addresses, one internet.
Why a device can have more than one kind of address, and why one route may work when another does not.
Two address families
An IP address helps a network deliver packets to a destination. IPv4 and IPv6 are different address families. IPv4 addresses are often written as four numbers separated by dots. IPv6 uses a longer hexadecimal notation with colons, and can shorten consecutive groups of zeroes.
A device can use both at the same time. This is commonly called dual stack. It does not mean every destination is reachable through both families, or that the two journeys use identical equipment along the way.
A name can have several answers
A domain name can resolve to IPv4 addresses through DNS A records and IPv6 addresses through AAAA records. Applications decide which available address to try. Many use connection strategies that reduce the wait when one family is slow or unavailable.
The existence of an IPv6 address alone does not prove a complete working connection. The device also needs an appropriate route, and the destination needs to accept the traffic. The same principle applies to IPv4.
Separate paths, separate symptoms
If a site works through IPv6 but fails through IPv4, it does not automatically mean the site itself is broken. Routing, filtering and connectivity can differ between the two paths. Testing both separately is useful because it narrows the problem without immediately changing unrelated settings.
Keep a note of the destination, the address family and the exact failure. A DNS error, a connection timeout and a rejected login point to different stages of the journey.
Keep the conclusion modest
Successful access over one path is evidence that that path works. It does not tell you everything about the other one. A careful comparison is more useful than treating IPv6 as universally faster or IPv4 as universally more compatible.
For everyday use, the goal is straightforward: let the device use working paths and make failures easy to identify.