I’m skeptical that Koorde is the right approach. It’s main advantage over Kademlia is that nodes only need constant size state (i.e. number of routing table entries) as a function of the size of the network, as opposed to Kademlia’s O(log(n)) state. The size of the routing table isn’t a limiting factor for us at the moment.
I think we should be more focused on robustness and request latency for now; see this discussion