REPAIR – REsilient Protocols And Internet Routing

The Internet is being adapted as a transport medium for a wide range of communication services, and the consequences of malfunctions in the infrastructure are severe. Current resilience mechanisms in the Internet are not designed to meet the stringent requirements of new interactive services like telephony and online gaming.

REPAIR will develop methods for increasing network resilience against failures and other deviations from normal operation. Within this context, several important topics will be explored, including

• Proactive local recovery. As the Internet takes an increasingly central role in our communications infrastructure, the slow convergence of routing protocols after a network failure becomes a growing problem. Traditional routing convergence is slow because it is reactive and global. It reacts to a failure after it has happened, and it involves all the routers in the domain. We are developing recovery mechanisms that are proactive and local. Such mechanisms can give very fast response to network failures without compromising stability.

• Adaptive multipath routing. To increase the resilience against failures and unexpected traffic patterns, we believe that a routing protocol should have two important properties. First, it should be able to use more than one path towards each destination. This is important both for recovery and load balancing purposes. Second, it should be adaptive to changing traffic conditions. A major challenge for load adaptive routing protocols is to avoid instabilities in the form of oscillations.

• Interdomain scalability. The scalability of interdomain routing seems to be a growing concern in the networking community. Scalability is an issue in two different respects: increasing routing table size, and an increasing number of
  update messages (churn).  We want to investigate how different aspects like the policy routing used in the Internet and the topological properties of the Internet graph influence scalability. Such understanding important input when designing new routing architectures: without a clear understanding of the scalability challenges that are facing
  interdomain routing, any proposed solution will be a shot in the dark.