Strange latency issue (no its not a congested exchange)

techead

Honorary Master
Joined
Apr 11, 2008
Messages
12,185
Reaction score
48
Location
by the mountain
unnamed.png

Pic above is of ZERO activity. Unfortunately I dont have a pic of activity, but its NOT pretty. Goes up around 300/400/500ms even if you playing an online game. Making it impossible to play actually.

NOTE : Line can download at full speed, pings are REALLY BAD then... :/

Symptom : Ping is fairly stable when no activity on the line. Its not great though. As soon as you have any form of activity on the line, and Im talking 10% of line, the ping goes to cr@p

His SNR values are Attenuation are actually really good. So I dont get it.

I have ruled out ISP, does the same thing on various backbones.

Could it be a bad connection somewhere?
 
Last edited:
View attachment 89041

Pic above is of ZERO activity. Unfortunately I dont have a pic of activity, but its NOT pretty. Goes up around 300/400/500ms even if you playing an online game. Making it impossible to play actually.

NOTE : Line can download at full speed, pings are REALLY BAD then... :/

Symptom : Ping is fairly stable when no activity on the line. Its not great though. As soon as you have any form of activity on the line, and Im talking 10% of line, the ping goes to cr@p

His SNR values are Attenuation are actually really good. So I dont get it.

I have ruled out ISP, does the same thing on various backbones.

Could it be a bad connection somewhere?

Are you running this ping from a wireless device?
 
I had the same kind of issue, logged a call with Telkom and the "re-configured" my port and all issues disappeared. I think that my port got swopped with someone else, but at least my issues were gone.

My resting ping was anything between 2500ms and 6500ms to the first node and after the re-configure it dropped down to 4ms.
 
Are you running this ping from a wireless device?

Initially I was running a ping from wireless yes, sitting about 3m from the router. I clicked, and plugged in a cable. Now Im not certain if this ping was still from wi-fi or from the cable. Regardless, the same issue remains... that I can assure you. As I spent the entire evening testing NOT using wi-fi
 
I had the same kind of issue, logged a call with Telkom and the "re-configured" my port and all issues disappeared. I think that my port got swopped with someone else, but at least my issues were gone.

My resting ping was anything between 2500ms and 6500ms to the first node and after the re-configure it dropped down to 4ms.

This is a different issue. Sounds more like congestion on the first hop. I dont have this. thanks for the reply though.

I will log it though , hopefully Telkom can work some magic
 
not enough information to help you.

that ping is pretty decent for pinging something in the UK. i would expect that ping to go higher with activity.

nothing strange
 
not enough information to help you.

that ping is pretty decent for pinging something in the UK. i would expect that ping to go higher with activity.

nothing strange

you joking right? :rolleyes: If you have a line that can do 200 KB/sec and you raping the line 100% then ofc your ping will go to hell.

If you have activity in the 20 KB/sec range and pings increase to 300/400/500 then there is something WRONG.

Im not debating whether something IS wrong. I know something IS wrong, Im merely looking at fellow forum members if they have seen something similar and what they did to rectify the situation.

EDIT : oh btw, THIS is what a ping to the UK should look like.... Rock stable on an inactive line with a ms either way here and there

unnamed (1).png
 
Last edited:
you are comparing latency responses from 2 different IP addresses.

if I ping those 2 different IP addresses I also get a difference in response times.

I don't know the route your connection takes to the destination IP's or what devices are at the end of the IP's. ICMP is at the bottom of the hierarchy for a lot of devices.


however, all that aside, a physical fault can cause latency problems when it starts transferring data. from your NIC to the cable to the router/modem to the POTS filter or at the telkom side of things. etc.


have you logged a fault with Telkom explaining that your latency crashes when you transfer data over your line?

have you checked your SNR and attenuation during idle and during line usage, it may change, which it shouldn't

what does a tracert show during idle and usage.


as I said more information is needed
 
you are comparing latency responses from 2 different IP addresses.

if I ping those 2 different IP addresses I also get a difference in response times.

Thats not really the point. Im comparing consistency.

I don't know the route your connection takes to the destination IP's or what devices are at the end of the IP's. ICMP is at the bottom of the hierarchy for a lot of devices.


however, all that aside, a physical fault can cause latency problems when it starts transferring data. from your NIC to the cable to the router/modem to the POTS filter or at the telkom side of things. etc.

yup, im working my way through all the possibilities. Tricky coz its not my line.


have you logged a fault with Telkom explaining that your latency crashes when you transfer data over your line?

I have through unofficial channels, yes. A friend of mine is looking into it.

have you checked your SNR and attenuation during idle and during line usage, it may change, which it shouldn't

Will look asap , thanks for this

what does a tracert show during idle and usage.

Will look asap, thanks for this
 
just remembered, one of the wires at your RJ45 plugs could have pulled out/loosened ever so slightly, it would also cause this.

vacuum cleaners or brooms and feet hooking on cables cause this to happen..
 
just remembered, one of the wires at your RJ45 plugs could have pulled out/loosened ever so slightly, it would also cause this.

vacuum cleaners or brooms and feet hooking on cables cause this to happen..

I assume you mean the RJ11 connector?

re RJ45 I got a bunch of network cables, so I did try a couple of different ones.
 
I assume you mean the RJ11 connector?

re RJ45 I got a bunch of network cables, so I did try a couple of different ones.



yeah in my experience if the rj11 plug/connection is damaged you will see it in your connection stats, but there's always the possibility...

if you've tested with other network cables, cool.

the default routine for testing is:

unplug all devices from other telephone sockets
plug a single device into your modem/router with a short cable
connect the modem/router to the main/master/first phone socket of the premises
make sure you use a pots filter on the wall socket side of the rj11 cable, not the modem side.

move your pc if needed...

if the problem still persists, try another filter, other cables, other modem/router(if available) or another NIC(if possible)


if the problem still persists, you have covered all your bases, and its over to the copper provider.


ages ago I have come across warez that activate as soon as you use your connection. low data but high packet spam. chances of this is low though..


something els to try

get drtcp to set your NIC MTU size to 1458, doubt it would help but worth a try.
 
current line stats while IDLE

Attenuation : 22.5 db
SNR : 24.2 db

line stats when downloading via laptop and streaming youtube on PC

Attenuation : 22.5 db
SNR : 24.1 db

very good line stats. Way better than mine at home.
 
Last edited:
Hmmm,more than 1 Switch/hub involved here?

Try pinging from 1 device to another on the same network,would like to see if the latency is only outgoing or internal
 
my telkom sauce came back to me... he logged a fault on the line. technician dispatched!

can anyone interpret this?

image005.jpg
 
Misc. Business Line (Under freqeuently usage)

Pinging 212.58.244.69 with 32 bytes of data:
Reply from 212.58.244.69: bytes=32 time=192ms TTL=53
Reply from 212.58.244.69: bytes=32 time=257ms TTL=53
Reply from 212.58.244.69: bytes=32 time=234ms TTL=53
Reply from 212.58.244.69: bytes=32 time=205ms TTL=53
Reply from 212.58.244.69: bytes=32 time=238ms TTL=53
Reply from 212.58.244.69: bytes=32 time=189ms TTL=53
Reply from 212.58.244.69: bytes=32 time=274ms TTL=53
Reply from 212.58.244.69: bytes=32 time=230ms TTL=53
Reply from 212.58.244.69: bytes=32 time=290ms TTL=53
Reply from 212.58.244.69: bytes=32 time=196ms TTL=53
Reply from 212.58.244.69: bytes=32 time=269ms TTL=53
Reply from 212.58.244.69: bytes=32 time=190ms TTL=53
Reply from 212.58.244.69: bytes=32 time=189ms TTL=53
Reply from 212.58.244.69: bytes=32 time=279ms TTL=53
Reply from 212.58.244.69: bytes=32 time=325ms TTL=53
Reply from 212.58.244.69: bytes=32 time=192ms TTL=53
Reply from 212.58.244.69: bytes=32 time=191ms TTL=53
Reply from 212.58.244.69: bytes=32 time=222ms TTL=53
Reply from 212.58.244.69: bytes=32 time=239ms TTL=53

190ms is normal to the UK (Give or take 30ms).

If you want more info about where the ping is going up, run a traceroute.

tracert 212.58.244.69

Tracing route to bbc-vip114.telhc.bbc.co.uk [212.58.244.69]
over a maximum of 30 hops:

1 9 ms 3 ms 4 ms *Removed*
2 10 ms 5 ms 4 ms *Removed*
3 121 ms 53 ms 55 ms *Removed*
4 28 ms 26 ms 20 ms cdsl1-rba-vl144.isdsl.net [196.35.63.65]
5 103 ms 97 ms 132 ms core1-pkl-tg0-7-0-0.ip.isnet.net [168.209.1.162]
6 206 ms 199 ms 206 ms 168.209.201.97
7 308 ms 222 ms 250 ms bbc-linx.pr01.thdow.bbc.co.uk [195.66.224.103]
8 * * * Request timed out.
9 * * * Request timed out.
10 346 ms 251 ms 252 ms ae0.er01.telhc.bbc.co.uk [132.185.254.109]
11 196 ms 191 ms 245 ms 132.185.255.149
12 278 ms * 321 ms bbc-vip114.telhc.bbc.co.uk [212.58.244.69]

Then ping the hops that seem higher than you think they should be :p
 
Last edited:
Top
Sign up to the MyBroadband newsletter
X