Brief substantive summary of the workshop and presentation of the main issues that were raised during the discussions
The workshop aimed at exploring the technical, policy, legal and law enforcement implications of the implementation of Carrier Grade Network Address Translation in the Internet. The workshop was organized as a round table with key speakers who have been working at multiple levels in the Internet. Significant time was set aside for interaction between all participants in the workshop. Over 45 representatives of various stakeholder groups attended the session.
The session began with a technical introduction to what Carrier Grade Network Address Translation (CGN) is and why Internet Service Providers are motivated to implement it. The implications of IPv4 address exhaustion were explored and statistics about both IPv4 addressing and IPv6 addressing were presented. Few users are aware of whether they are behind CGN technology and even fewer are aware of the implications. After reviewing the current state of IPv4 addressing in the contemporary Internet, the workshop explored some of the effects of implementation of the technology. In particular, a discussion took place where an itemized list of things that break in the presence of CGN technology was presented.
Panelists began with a discussion of the technical implications of CGN. In one study, more than 50% of traffic examined appeared to transit through CGNs. This implies that CGN, as a technology, is much more widely deployed than many people recognize. The implications of this on the Internet’s architecture was examined with special emphasis placed on the early design of the Internet where intelligence was placed at the edges of the network and the core of the network was responsible for nothing more than transit.
A short discussion regarding CGN technology use in mobile networks. It was noted by some of the panelists that mobile networking is quite different from traditional Internet provision in that there is significant control over what gets attached to the network and there are seldom applications that rely on inbound connections to edge devices. As a result, it was noted that CGN technology was widely used in mobile networks – precisely because those networks and the devices attached are largely heterogeneous.
Panelists noted the implications on particular protocols, especially peer-to-peer protocols. There was a lively discussion on the economic likelihood of organizations abandoning CGN technology once it was deployed. An economic analysis of the long-term effects of CGN implementation was then presented. Panelists generally agreed that, once an Internet Service Provider moved to implement CGN technology, there was little economic incentive to move away from it.
The discussion was then opened up to the workshop as a whole. The implications of CGN in developing economies was explored with a discussion of the prevalence of CGN technology in those settings. Once again, the panelists and the workshop as a whole discovered that CGN technology is more widely deployed than many would have expected.
Conclusions drawn from the workshop and possible follow up actions
There was near consensus that CGNs are more widely deployed than most people realize. It was also largely agreed that CGNs have significant implications on end users and protocols on the Internet. However, it was also agreed that few users of the Internet realize the implications of deployment of this technology. Those implications were shown to cut across countries and continents – the problems are the same for developed and developing economies. The workshop concluded with a call for further efforts at raising the level of information available about the implications of CGNs.