Latency on ADSL Line - Eldoraigne X6

[chillnet]

Hi Guys,

Telkom recently upgraded my line speed from 2M to 4M. The sad thing is that my line is now more degraded than before. I am not as clued up with the SNR ect, but I think this is bad!

Update Counter : 41403
Modulation : G.Dmt
Annex Mode : Annex A/L
Line State : up
Lan Tx : 4569262
Lan Rx : 2353802
ADSL Tx : 264162
ADSL Rx : 315198
CRC Down : 1
CRC Up : 23
FEC Down : 0
FEC Up : 0
HEC Down : 23
HEC Up : 1
SNR Up : 14.0
SNR Down : 30.0
Line Attenuation Up : 15.0
Line Attenuation Down : 29.5
Data Rate Up : 512
Data Rate Down : 4096

I also did a traceroute this morning to one of our servers in Hetzner's Midrand DC:

traceroute placementpartner.co.za
traceroute to placementpartner.co.za (41.203.24.180), 64 hops max, 52 byte packets
1 my.router (192.168.1.1) 3.371 ms 1.767 ms 1.608 ms
2 41.183.22.1 (41.183.22.1) 908.748 ms 920.939 ms 716.622 ms
3 41.183.141.25 (41.183.141.25) 717.517 ms 914.739 ms 820.038 ms
4 41.183.132.68 (41.183.132.68) 818.437 ms 734.279 ms 716.105 ms
5 41.183.132.82 (41.183.132.82) 713.633 ms 793.473 ms 843.947 ms
6 196.37.220.21 (196.37.220.21) 843.177 ms 820.687 ms 1023.315 ms
7 197.96.148.74 (197.96.148.74) 1022.698 ms 920.773 ms 717.658 ms
8 * core.gp-cn-het-mee-1.to.gp-mr-vod-mee-1.dfa.p2p.10g.za.africainx.net (41.84.12.151) 528.266 ms 557.610 ms
9 41-66-132-242-f2.het001-cpe-1-to-gp-cn-het-mee-1.africainx.net (41.66.132.242) 561.866 ms 708.430 ms 618.555 ms
10 core-access-switch1.jnb1.host-h.net (197.189.193.1) 735.699 ms 679.497 ms 747.001 ms
11 row-access-switch1-row7-8.jnb1.host-h.net (197.189.193.39) 882.992 ms 609.467 ms 678.526 ms
12 firewall1.jnb2.host-h.net (41.72.136.68) 878.377 ms 807.752 ms 939.904 ms

And then to one of my servers in the UK:

traceroute chillnet.cc
traceroute to chillnet.cc (78.129.247.26), 64 hops max, 52 byte packets
1 my.router (192.168.1.1) 45.155 ms 2.130 ms 2.279 ms
2 41.183.22.1 (41.183.22.1) 918.743 ms 921.653 ms 722.277 ms
3 41.183.141.25 (41.183.141.25) 1219.493 ms 788.747 ms 754.390 ms
4 41.183.132.68 (41.183.132.68) 958.683 ms 771.550 ms 786.576 ms
5 41.183.132.82 (41.183.132.82) 920.957 ms 819.230 ms 714.159 ms
6 196.37.220.21 (196.37.220.21) 820.876 ms 690.239 ms 816.910 ms
7 ar2-rba-gi0-2.ip.isnet.net (168.209.86.249) 818.662 ms 716.591 ms 817.745 ms
8 core1b-pkl-te0-0-0-0.ip.isnet.net (196.26.0.62) 715.309 ms 714.784 ms 921.058 ms
9 168.209.201.89 (168.209.201.89) 1127.426 ms 883.094 ms 911.994 ms
10 te8-4.core1.thn.as20860.net (195.66.224.207) 1526.809 ms 1125.148 ms 1024.534 ms
11 593.net1.north.dc5.as20860.net (62.233.127.174) 1026.969 ms 1022.909 ms 1023.738 ms
12 87.117.211.46 (87.117.211.46) 1023.515 ms 1006.091 ms 1025.864 ms
13 * *

I've logged a fault with Telkom but not getting any feedback from them...

Any suggestions?

 

[pezulu]

Eldoraigne (EDO3) has been having latency issues, slow speeds and such since Friday last week.
Contacted TI and logged a fault with them.
Feedback yesterday was that the line has been retested and no faults were found.

Happened to meet a Telkom technician yesterday, and he stated that he suspects that the exchange is congested, as capacity has not been increased, yet everybody has been upgraded.

I am supposed to be on a 10Mbps line, but max speed available at the exchange is limited to 4Mbps.

Pings were running in the high 2k, download speed was about 0.45kbps, upload speed was about 0.25kbps.

Since this morning speeds and latency seems to be almost back to normal.

 

[JonesY]

Telkom recently upgraded my line speed from 2M to 4M. The sad thing is that my line is now more degraded than before

Packets that can't get through - on a congested DSLAM - have to be retransmitted, as do the packets they collide with. A higher sync profile on a congested trunk will only cause more collisions, resulting in more retransmissions and also a greater reduction in throughput.

Have your sync speed decreased again or downgrade your plan altogether; you'll save cash and it may also improve overall performance.

 

[Shard]

Packets that can't get through - on a congested DSLAM - have to be retransmitted, as do the packets they collide with. A higher sync profile on a congested trunk will only cause more collisions, resulting in more retransmissions and also a greater reduction in throughput.

Have your sync speed decreased again or downgrade your plan altogether; you'll save cash and it may also improve overall performance.

What he said, downgrade always seems to sort it out, you'll have to go down to 1Mbps though, and get 2Mbps next year with upgrade rollout.

 

[JonesY]

What he said

Because only men know a bit about congestion.

 

[techead]

snip

welcome to the world of a congested exchange