Project Summary: NeTS: Small: Modeling IPv6 Adoption: A Measurement-driven Computational Approach

Broad industry awareness of IPv4 address scarcity has driven widespread support for IPv6 in most modern operating systems and network equipment, but even with years of encouragement from regional address registries (RIRs), and government mandates, the actual uptake of IPv6 has been disappointing. A vast ecosystem of software applications and devices still do not support IPv6, and many technologies designed to support the transition add complexity and reduce performance and reliability. Further complicating the situation, the RIR communities have approved policies that could have the unintended consequence of halting the adoption of IPv6 altogether. Understandably then, there is a blinding lack of consensus on whether IPv6 will ever be widely adopted by ISPs, leaving the broader Internet ecosystem with tremendous uncertainty and risk in planning for the future. While some companies have avoided IPv6 deployment, others have undertaken valiant efforts to be early adopters of IPv6, risking loss of their investment should the transition ultimately fail. Given these daunting circumstances, even skeptics of our ability to comprehend the dynamics of such a large and complex network as the Internet must consider the utility of measurements to characterize factors affecting IPv6 deployment, and a model to inform the actions of stakeholders in the transition to IPv6. We propose a two-phase project to measure and computationally model the adoption of IPv6. In the first phase we will extensively measure two phenomena in the current IP addressing ecosystem that may either cause and/or reflect IPv6 deployment decisions: market-based transfers of IPv4 address blocks and deployment of Carrier Grade NATs (CGNs). The measurement phase will deliver detailed characterizations of these emerging phenomena, and empirically ground the second phase of our project: developing and applying a computational model that captures decision processes of key stakeholders, accounting for the impacts of costs, performance, address markets, the availability of alternate solutions (e.g., CGNs), and geographic/political constraints on IPv6 deployment decisions. We will use the model to explore the influence of several known factors, individually and in combination, and perform predictive modeling of the impact of those factors and decisions of Internet stakeholders on the state of overall IPv6 adoption. Intellectual merit: The collected measurement data and quantitative and predictive nature of the modeling effort will enable transformative research on the most important architectural transition thus far of the Internet. Existing models for studying IPv6 deployment ignore the reality that in most organizations transition will occur incrementally — …