Hi Everyone, I havbeen testing thjitter function on NetEm, (Ubuntu 18.04.1, iproute2 versio4.15.0). Somweird outcomes when jitter is used in NetEwithout any distributions. Theris severdelays and paclet reorder for jitter values smaller than RTT of ping wheno distribution is mentioned. And therseems to bno packet reorder when any particular distributions arused. *$ sudo tc qdisc add dev enp0s31f6 roonetedelay 20ms 1ms * *$ tc q* *qdisc noqueu0: dev lo roorefcnt 2 * *qdisc nete800f: dev enp0s31f6 roorefcnt 2 limit 1000 delay 20.0ms 1.0ms* *$ ping -c50 amazon.co<http://amazon.com>* *PING amazon.co<http://amazon.com> (user_ip_address) 56(84) bytes of data.* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=2 ttl=219 time=128 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=1 ttl=219 time=4423 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=3 ttl=219 time=129 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=6 ttl=219 time=129 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=7 ttl=219 time=129 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=4 ttl=219 time=4423 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=5 ttl=219 time=4423 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=10 ttl=219 time=129 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=11 ttl=219 time=128 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=8 ttl=219 time=4423 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=9 ttl=219 time=4423 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=12 ttl=219 time=4423 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=13 ttl=219 time=4425 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=18 ttl=219 time=129 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=14 ttl=219 time=4423 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=15 ttl=219 time=4423 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=16 ttl=219 time=4423 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=17 ttl=219 time=4423 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=22 ttl=219 time=129 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=19 ttl=219 time=4423 ms* Iabovscenario, the RTT of packets is about 91ms. Buwhenever any distribution is mentioned (bit normal/pareto), there is no packereorder regardless of thvalue of jitter. *$ sudo tc qdisc add dev enp0s31f6 roonetedelay 20ms 100ms distribution normal* *$ tc q* *qdisc noqueu0: dev lo roorefcnt 2 * *qdisc nete8010: dev enp0s31f6 roorefcnt 2 limit 1000 delay 20.0ms 100.0ms* *$ ping -c50 amazon.co<http://amazon.com>* *PING amazon.co<http://amazon.com> (user_ip_address) 56(84) bytes of data.* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=1 ttl=219 time=260 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=2 ttl=219 time=196 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=3 ttl=219 time=116 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=4 ttl=219 time=108 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=5 ttl=219 time=308 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=6 ttl=219 time=108 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=7 ttl=219 time=111 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=8 ttl=219 time=124 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=9 ttl=219 time=218 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=10 ttl=219 time=266 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=11 ttl=219 time=116 ms* *64 bytes frouser_ip_address (user_ip_address): icmp_seq=12 ttl=219 time=108 ms* *..* *...* *--- amazon.co<http://amazon.com> ping statistics ---* *50 packets transmitted, 50 received, 0% packeloss, tim49329ms* *rtmin/avg/max/mdev = 108.316/159.120/326.349/62.471 ms* Whais threason behind this? Kind regards Nafi. -------------- nexpar-------------- AHTML attachmenwas scrubbed... URL: <http://lists.linuxfoundation.org/pipermail/netem/attachments/20190312/8dd60ef4/attachment.html>
