There’s another interesting article from Geoff Huston (APNIC) published on CircleID, analysing claims that IPv6 can offer inferior performance compared to IPv4. Geoff has talked about this before, but APNIC Labs have now developed an online reporting tool showing the relative performances of IPv6 compared to IPv4 for nearly 150 countries and territories.
The tool features a clickable world map as well as tables for global regions, individual countries/territories and by AS number showing two different types of data. The first set of data compares round trip times (RTTs) for both IPv6 and IPv4, with the second set of data compares IPv6 connection robustness.
Some conclusions drawn are that where paths between the same two dual stack endpoints are examined, both IPv6 and IPv4 protocols as measured by TCP SYN round trip time, are roughly equivalent on average. IPv6 appears to be faster on average in Europe, Africa and Oceania, whereas IPv4 is faster than average in Asia and the Americas, although this is also very dependent on particular access networks.
As previously reported at RIPE 71, TCP connection performance is roughly comparable once a connection is established, but the failure rate for unicast IPv6 is 1.6% compared with just 0.2% for IPv4. However, this failure rate has decreased substantially since 2011, and even since last year where it stood around 4%. This suggests that connection reliability improves as more IPv6 is deployed, and particularly as there becomes less need to rely on transitional mechanisms.
At Deploy360, we continue to encourage the use of IPv6. Please take a look at our Start Here page to understand how you can get started with IPv6.