Juliusz Chroboczek <[email protected]> writes: >> I did not read mentioned thread, but note that here the function >> babel_find_source() iterates only through sources for one propagated >> prefix. > > Ah, I missed that. So yeah, no problem.
Ah, right. Was wondering why that would be a problem... Yeah, there's a FIB for the prefix lookup :) Hmm, benchmarking would be interesting; maybe I'll try to get a node attached to that Slovenian network and see what breaks :) >> I have one more general question. Is ETX cost algorithm really >> a reasonable idea? [...] It seems to me that in many cases it will lead >> to oscillations, because packet loss in wireless networks usually >> depends on traffic load. It seems to me that dynamic cost algorithms >> based on packet loss should at least use some hysteresis. > > You're right, we're doing hysteresis in babeld. Yeah, that ended up on my "nice to have" list, so didn't implement it in the first round. Related to that: Ondrej, I believe I saw you mention in another thread that the planned 2.0 release will be coming out soonish. Does that mean there will be dual-stack capabilities? That, and source-specific routing, are the two main things missing from Bird to get this Babel implementation to feature parity with babeld... I'll be happy to look into that at some point in the future (after you're finished going over this version with that fine-toothed comb of yours) :) > I'll be deploying one or two nodes with Toke's code as soon as I redo > our bench network (migrating from AHCP to HNCP), we'll see whether it > introduces any pathologies. Yay! Will look forward to the torrent of bug reports ;) -Toke
