Oh, that would have been 100 ms unloaded. On 12/1/15 10:08 AM, Dave Taht wrote: > What RTT was fig 7 in the codel paper? (the one that showed codel's > reaction to bandwidth changes?) > http://queue.acm.org/detail.cfm?id=2209336 > > I am amusing myself by trying to create an automated string of flent > tests to duplicate those results, now that flent has gained tools that > can plot queue depth as well as ping, we have a testbed with fully > controllable RTT, etc. > > http://snapon.cs.kau.se/~d/nichols/basiccodel.png > > "To roughly emulate a (nominal) 100-Mbps Wi-Fi link subject to > degradation, we used a load of four FTPs and five Web connections per > second and changed link rates at 50-second intervals (over the 300 > simulated seconds), first dropping to 10 Mbps, then to 1 Mbps, then > jumping to 50 Mbps, dropping to 1 Mbps, and finally jumping back to > 100 Mbps. Buffer capacity is a single BDP (830 packets) for the > nominal rate. This scenario was repeated for CoDel, Tail Drop, and > RED. " > > -- > Dave Täht > Let's go make home routers and wifi faster! With better software! > https://www.gofundme.com/savewifi > _______________________________________________ > Codel mailing list > [email protected] > https://lists.bufferbloat.net/listinfo/codel >
_______________________________________________ Codel mailing list [email protected] https://lists.bufferbloat.net/listinfo/codel
