Empirical Performance of IPv6 vs. IPv4

 under a Dual-Stack Environment





 Measurement Setup

 Dual-Stack List


 Hop Count



 OS Dependence

 IPv6 Address


 IPv6 Tunnel




 Internal Access



Due to the practical difficulties in obtaining “large” blocks of new, unassigned IPv4 addresses, major organizations in the fast-growing markets of Asia and Europe, as well as mobile service providers worldwide, are under increasing pressure to migrate from the entrenched IPv4 standard to the emerging IPv6 one. As we continue to see increasing global-scale deployments of IPv6 networks, there has also been an increasing interest in measuring the performance of these IPv6 networks [1]-[7].

In this work, we want to characterize, from an end-user's perspective, the differences in actual networking performance as one switches from IPv4 to IPv6. Towards this end, we have conducted extensive measurements by sending probing traffic and initiating file downloads from our dual-stack IPv6/IPv4 testbed to over 2,000 dual-stack hosts worldwide.

Our study follows the comparative approach of [1], [5] but considerably broaden the scope of the previous studies by covering more than 2000 dual-stack end-hosts worldwide. In addition, by comparing our results with previously published results, we can quantify the maturation of IPv6 in the past few years from a deployment and performance-enhancement perspective. In our work, we measure the network performance of IPv6 networks in terms of various network metrics like network connectivity, hop count, RTT, throughput, operating systems dependencies as well as the IPv6 address configuration latency. We also study the network performance of three major IPv6 tunnel brokers and compare them to native IPv6 services.