Following on from last year's measurements after World IPv6 Day (see an earlier post on CircleID: Comparing IPv4 and IPv6 Performance), the RIPE NCC carried out active measurements on World IPv6 Launch on 6 June 2012. These measurements included latency measurements both on IPv4 and IPv6 from our vantage points to selected hostnames of World IPv6 Launch participants and other dual-stacked parties. We used these measurements to determine the performance of IPv4 versus IPv6 connections.
The figure below shows a histogram of all relative IPv4 versus IPv6 performance data points collected during World IPv6 Launch. A single data point consists of the relative performance of IPv4 and IPv6 between one of our vantage points and a World IPv6 Launch participant over a15 minute time frame.
The peak in the middle of this histogram shows that the IPv4 and IPv6 performance that we measured is roughly similar. Comparable IPv4 and IPv6 performance can be seen as an indication of mature deployment of both IPv4 and IPv6 in a network.
The fact that the bell-shaped curve is a little "fatter" on the IPv4 side, indicates that when performance differs, it is usually IPv4 that is faster than IPv6.
For end-hosts, being dual-stacked means there are two methods (IPv4 and IPv6) to get packets to another dual-stacked end-host, so they are not only relying on IPv4. This year we saw that IPv6 was already faster in roughly 10% of the cases, and with further maturation of the IPv6 network this is only expected to rise.
So being dual-stacked means two chances for better performance!