Testing a Flare-like routing implementation on 2500 AWS nodes

However, he mentioned that the Flare paper needed more work as it only included a series of simulations of various topologies/parameters which are then extrapolated to larger network sizes. The next logical step would be to deploy a proto-implementation within a live testbed with real latencies and preferential attachment. He was glad that they went ahead to gauge the real-word feasibility of their scheme.In response to this, Gary suggested using modules in Linux iptables that allow simulating packet loss and latencies. Although any simulation is not as good as a real-world test, it can perform a lot of negative testing in the cloud with faster test/debug/fix cycle times than in a real-world deployment. Gary also offered his help with iptables config if there was interest.