Monday, April 20, 2009

Building a better Internet - the Pouzin Society

[Here is a group worth following – especially some of their white papers with contributions from luminaries such as John Day, David Meyer, Mike O’Dell – BSA]

The Pouzin Society, named after Louis Pouzin, the inventor of datagrams and connectionless networking, announces its initial organizing meeting. The society’s purpose is to provide a forum for developing viable solutions to the current Internet architecture crisis.

About 15 years ago, it became clear that IPv4 was reaching its limits, and the IETF responded by creating IPv6. In 2006 came the tacit admission that there continue to be fundamental scaling problems in the Internet routing architecture which would only be exacerbated by IPv6, and that Moore's Law could not save us this time. Several solutions were proposed, all based on revising IPv6 addressing using the concept of a locator/identifier split. Work has proceeded diligently, but a few months ago, it became clear that not only was this approach fatally flawed, but by implication, so was IP, or any variation of it. Academic efforts, beginning with NewArch and continuing with FIND and GENI are no closer to finding a solution than we were a decade ago.

In the meantime, “Patterns in Network Architecture” has appeared, describing a simple new architecture that not only solves existing problems but also predicts capabilities as yet unconsidered.

Initial meetings will be held in conjunction with FutureNet in Boston, May 4 - 7. There will be a one day organizing meeting on May 4 to discuss collaboration and next steps. On May 6 and 7 there will be a working meeting at Boston University on the specific topic of the current addressing crisis. There will be considerable work refining architectural details, but the central goal of the effort is to form a group that builds implementations of this new network architecture to evaluate its scalability, security, and other pertinent characteristics.