Transport Architectures for Very Resilient Internet (TAVRI)


The current Internet is strictly dependent from the same protocol (BGP), largely deployed in millions of routers, yet being extremely weak against a number of attacks, the most frequent and dangerous ones being routing attacks. Additionally, the single way to achieve a reliable interconnection at the Internet edges under BGP is to intensively tweak its parameters, operations that undermine the overall Internet resiliency and poses important scalability problems (huge routing tables and routing instabilities).

The TAVRI project proposes a rather simple Internet routing principle yet extremely revolutionary that can unblock the development of advanced services in a secure and resilient future Internet. Our will is to define a solution that is totally transparent to BGP while allowing its partial replacement, especially at the transit core of the Internet.

The framework is that of Transit-Edge hierarchical routing, through which we pass from a flat Internet routing to a 2-level hierarchical routing, between edge networks from on hand and between transit network from the other hand. Transit-Edge routing separation represents the key brick toward a far more resilient and secure Internet, tackling the resiliency and security aspects from the more essential direction that underfund the Internet infrastructure: the routing principle. The idea is to apply basic mailing principles to Internet routing: as a mailbox can potentially leave the post network at many post offices, a packet can reach the destination network passing through many routing locators; as routing among post offices is much more resilient and fast than from the post office to the destination address, the traffic can be better aggregated and engineered in the Internet transit core. The TAVRI project work towards the definition of resilient protocols and traffic engineering solutions appropriate to this future context.