On Fri, Aug 1, 2008 at 10:39 PM, Theodore Hong <theodore.hong at cl.cam.ac.uk> wrote: > Ian Clarke <ian.clarke at gmail.com> wrote: >> On Thu, Jul 31, 2008 at 5:22 PM, Matthew Toseland >> <toad at amphibian.dyndns.org> wrote: >>> On Thursday 31 July 2008 17:42, Florent Daigni?re wrote: >>> IMHO the problem is location churn. The solution is for opennet nodes not to >>> swap. But we need simulations before we can deploy that. Which Vive promised >>> to do... >> >> Why would opennet nodes possibly benefit from swapping? That works in >> opposition to the edge selection algorithm we use for opennet. I'd >> deploy that fix immediately, with or without a simulation. We can >> then follow-up with a simulation to confirm it. > > Hm, I would be wary of making algorithm changes without having some > evidence first to show that they'll work, or at least not damage the > current network. > > theo
Currently, putting a openet-only node offline for a few hours would result in location change -- this is really bad. I can't find where is the opennet edge selection code is -- can anyone confirm it have been implemented?