WebAfrica - Evening Latency issues (from Cape Town)

N

NightShift

Expert Member
Joined
Oct 8, 2009
Messages
1,952
Reaction score
1,404
2 nights in a row now gaming has been a no-go because of terrible latency aka lag.

I did a tracert to a couple of destinations, but unity3d is in the example below:

C:\Users\Me\>tracert unity3d.com

Tracing route to unity3d.com [169.54.221.59]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms 192.168.1.1
2 * * * Request timed out.
3 47 ms 41 ms 40 ms 196.38.75.122
4 63 ms 41 ms 41 ms 196.38.77.88
5 42 ms 42 ms 42 ms mi-za-cpt-p7-te0-7-0-0.ip.isnet.net [168.209.2.12]
6 545 ms 468 ms 644 ms core1b-dock-ten-0-4-0-0sub8.ip.isnet.net [168.209.246.65]
7 583 ms 718 ms 573 ms bbr01.lon01.networklayer.com [195.66.225.183]
8 187 ms 186 ms * ae5.cbs01.tg01.lon01.networklayer.com [169.45.18.12]
9 * * * Request timed out.
10 771 ms * 617 ms ae0.cbs01.eq01.chi01.networklayer.com [50.97.17.48]
11 727 ms 728 ms 806 ms ae2.cbs02.dr01.dal04.networklayer.com [169.45.18.4]
12 * * * Request timed out.
13 641 ms 676 ms 924 ms ae28.dar01.dal09.networklayer.com [50.97.17.221]
14 945 ms 667 ms 652 ms 99.76.17c6.ip4.static.sl-reverse.com [198.23.118.153]
15 306 ms 310 ms 309 ms web-cms-us-tx-9.unity.com [169.54.221.59]
16 320 ms 301 ms 294 ms web-cms-us-tx-9.unity.com [169.54.221.59]

Trace complete.

I cant tell where that IP address where the lag kicks in is located - Joburg or London. Is something going on, and is anyone aware of it? I can't seem to find a reason for this! And I surely cant be alone :/
 
I'm having slight packet loss in my CS:GO on local servers, but like I have a 600ms ping to my office.

Here's how my traceroute to unity3d looks on my WebAfrica business capped account:
Code: 
Tracing route to unity3d.com [169.54.221.58]
over a maximum of 30 hops:

  1    <1 ms    <1 ms    <1 ms  router.local [192.168.11.1]
  2     *        *        *     Request timed out.
  3    25 ms    24 ms    24 ms  196.38.75.121
  4    82 ms    82 ms    82 ms  196.38.75.122
  5    40 ms    41 ms    39 ms  196.38.77.88
  6    85 ms    90 ms    97 ms  mi-za-cpt-p7-te0-7-0-0.ip.isnet.net [168.209.2.12]
  7   195 ms   197 ms   187 ms  168.209.201.28
  8     *        *        *     Request timed out.
  9     *        *        *     Request timed out.
 10   232 ms     *        *     2e.13.2da9.ip4.static.sl-reverse.com [169.45.19.46]
 11     *        *        *     Request timed out.
 12   354 ms   351 ms   349 ms  ae2.cbs02.dr01.dal04.networklayer.com [169.45.18.4]
 13     *        *        *     Request timed out.
 14   287 ms   287 ms   291 ms  ae28.dar01.dal09.networklayer.com [50.97.17.221]
 15   963 ms   899 ms   918 ms  9d.76.17c6.ip4.static.sl-reverse.com [198.23.118.157]
 16   283 ms   285 ms   287 ms  web-cms-us-tx-9.unity.com [169.54.221.58]
 17   286 ms   282 ms   280 ms  web-cms-us-tx-9.unity.com [169.54.221.58]

Trace complete.

And here is how it looks towards mybroadband.co.za:
Code: 
Tracing route to mybroadband.co.za [104.20.10.169]
over a maximum of 30 hops:

  1    <1 ms    <1 ms    <1 ms  router.chris.local [192.168.11.1]
  2     *        *        *     Request timed out.
  3   167 ms    24 ms    26 ms  196.38.75.121
  4    75 ms    72 ms    72 ms  196.38.75.122
  5    42 ms    42 ms    43 ms  196.38.77.88
  6   545 ms   439 ms   416 ms  mi-za-cpt-p8-te0-0-0-0.ip.isnet.net [168.209.6.13]
  7    38 ms    39 ms    39 ms  168.209.6.130
  8   109 ms   114 ms   115 ms  196-10-140-198.ixp.capetown [196.10.140.198]
  9   146 ms   145 ms   132 ms  104.20.10.169

Trace complete.
 
This has been going on for 2 weeks now. I've notified Openserve, Webafrica and IS about this 2 weeks ago. Since the initial issues started I have begged multiple parties at these 3 places for information as to what the actual ef is going on. But hey so far they've all just given the finger and clearly aren't bothered to either communicate or bother to fix the issue....

Is your internet working, do websites load? Great your internet is just fine - This seems to be their attitude. They don't care if your latency is 5000 or 500 or 50, its all the same, your internet is working right? Great.
 
Exactly the same problem in PE in the evenings from around 6pm.

Trace Routes timeout to many websites which simply won't resolve or are slower than dial-up from the 90's.

In the morning it's fine again.

For ****s and giggles: News24 wouldn't resolve or load at all last night, and Facebook wouldn't load any images.

Started around 2 weeks ago with the Openserve F***-up or whatever it was.

Need to post this in the main WebAfrica support thread so they can sort it out.

This is BS now.
 
This issue appears to be worse on ISPs that use Internet Solutions as upstream provider. E.g WebAfrica, CrystalWeb and several others. Would also appear to affect primarily Western Cape areas.
 
hjlinde said:
This issue appears to be worse on ISPs that use Internet Solutions as upstream provider. E.g WebAfrica, CrystalWeb and several others. Would also appear to affect primarily Western Cape areas.
Click to expand...

I raised a ticket via the webafrica website, but it looks like it's been closed, and they asked me to do a tracert to webafrica.co.za, which I did, and to which I've had no response. Honestly, the people on these support desks sometimes....

From the tracert, it looks like the issue is between SA and London, and potentially in London. The IP address in question is Dimension Data owned according to some geo-location thing I checked, so the Internet Solutions association checks out.

It's likely nothing that the fools in SA can sort out, but it would be nice if they tried ffs.
 
Markd said:
I raised a ticket via the webafrica website, but it looks like it's been closed, and they asked me to do a tracert to webafrica.co.za, which I did, and to which I've had no response. Honestly, the people on these support desks sometimes....

From the tracert, it looks like the issue is between SA and London, and potentially in London. The IP address in question is Dimension Data owned according to some geo-location thing I checked, so the Internet Solutions association checks out.

It's likely nothing that the fools in SA can sort out, but it would be nice if they tried ffs.
Click to expand...

I left CrystalWeb around 2 years ago for a similar IS related peering issue that was never fixed.

I recall spending hours on pingplotter proving the problems in PE and CT with detailed traces. Didn't make an iota of a difference.

It appears WebAfrica is going down the same road.

Sad but true - the skills needed to diagnose and fix this do not exist within the ISP itself, and they couldn't care about a few customers who notice the severe drop off in performance.
 
And magically at around 10pm it's back to normal.

Couldn't even resolve some local news and tech sites with traceroute timeouts between 8ish and 10.
 
thewusman said:
And magically at around 10pm it's back to normal.

Couldn't even resolve some local news and tech sites with traceroute timeouts between 8ish and 10.
Click to expand...
Same here. Had 1-4% packet loss in CS:GO to local server between 20:00 - 22:00 and 20ms higher ping than usual.

The game was unplayable and now between 22:30 - 00:00 I had no issues.

Here is a traceroute at 21:00:
Code: 
Tracing route to 155.133.238.162 over a maximum of 30 hops

1 <1 ms <1 ms <1 ms router.local [192.168.11.1]
2 * * * Request timed out.
3 52 ms 60 ms 64 ms 196.38.75.121
4 25 ms 27 ms 26 ms 196.38.75.122
5 63 ms 62 ms 56 ms 196.38.77.88
6 70 ms 72 ms 56 ms mi-za-cpt-p7-te0-7-0-0.ip.isnet.net [168.209.2.12]
7 67 ms 68 ms 66 ms 168.209.100.101
8 75 ms 80 ms 81 ms 196.26.0.130
9 476 ms 442 ms 429 ms 196 - 60 - 9 - 22.ixp.joburg [196.60.9.22]
10 64 ms 65 ms 63 ms 192.168.210.55
11 83 ms 87 ms 82 ms 155.133.238.162

Trace complete.

vs one at midnight:
Code: 
Tracing route to 155.133.238.162 over a maximum of 30 hops

  1    <1 ms    <1 ms    <1 ms  router.local [192.168.11.1]
  2     *        *        *     Request timed out.
  3    25 ms    24 ms    26 ms  196.38.75.121
  4    25 ms    24 ms    25 ms  196.38.75.122
  5    25 ms    28 ms    24 ms  196.38.77.88
  6    25 ms    24 ms    25 ms  mi-za-cpt-p7-te0-7-0-0.ip.isnet.net [168.209.2.12]
  7    41 ms    40 ms    41 ms  168.209.100.101
  8    40 ms    40 ms    40 ms  196.26.0.130
  9    41 ms    40 ms    41 ms  196-60-9-21.ixp.joburg [196.60.9.21]
 10    41 ms    40 ms    41 ms  192.168.210.53
 11    40 ms    40 ms    41 ms  155.133.238.162

Trace complete.

My Axxess account had similar issues - and they're also using IS.
 
So a widespread IS backbone problem then ?

I am going to try some pingplotter graphs tonight.

Techncentral.co.za wouldn't even resolve last night during this time - it was worse than dial-up.

Something is fooked.
 
Hummercellc said:
My wholesale IS connection is perfectly fine. Maybe WebAfrica have not bought enough capacity from IS
Click to expand...

Initially I also suspected this but:

Netflix peak congestion times kick in earlier (way before 8pm); and

It’s affecting other IS backbone providers as well during similar times.

But Hey who knows. Could be anything at this stage all I do know is it’s awful.
 
Ok - maybe this really is a capacity issue.

8:30pm - netflix is crawling.

Latency up the pole.
 
hjlinde said:
This issue appears to be worse on ISPs that use Internet Solutions as upstream provider. E.g WebAfrica, CrystalWeb and several others. Would also appear to affect primarily Western Cape areas.
Click to expand...

Yup, if I switch to an Afrihost prepaid account I have all is fine..

Webafrica:
webafrica_netflix.jpg

Afrihost: (same Netflix server)
afrihost_netflix.jpg
 
Last edited:
Just go look at my pingplotter graph and stats from last night in the main WA thread. It's self explanatory.

IS Clusterf***K
 
Last edited:
treebsd said:
out of interest, which dns are you using; am asking since netflix is local on the IS network, you should not break out internationally.
Click to expand...

Good point, I switched to googles 8.8.8.8 a while back when there were dns issues - doesn't resolve the overall latency problems but should sort out Netflix for now, thanks.
 
Joker said:
Good point, I switched to googles 8.8.8.8 a while back when there were dns issues - doesn't resolve the overall latency problems but should sort out Netflix for now, thanks.
Click to expand...

k, that takes care of netflix.

give me more info on the latency issues you are experiencing. let me see how i can help.
 
Trace to www.news24.com:

Code: 
Tracing route to www.news24.com [41.86.110.200]
over a maximum of 30 hops:

  1     1 ms     1 ms     1 ms  router.asus.com [192.168.10.1]
  2     *        *        *     Request timed out.
  3     4 ms     5 ms     6 ms  196.38.75.121
  4     *       20 ms    21 ms  196.38.75.122
  5    23 ms    27 ms    25 ms  196.38.77.88
  6     5 ms     5 ms     5 ms  mi-za-cpt-p8-te0-0-0-0.ip.isnet.net [168.209.6.13]
[B]  7   419 ms   475 ms   665 ms  168.209.2.131[/B]
  8     6 ms     4 ms     5 ms  mweb-2.cinx.net.za [196.223.22.25]
  9     7 ms     7 ms     5 ms  197-84-7-54.cpt.mweb.co.za [197.84.7.54]
 10    13 ms    11 ms    10 ms  197-84-247-89.cpt.mweb.co.za [197.84.247.89]
 11     7 ms     6 ms     4 ms  OPTI-CPT-N1-Hosting.optinet.net [197.84.196.6]
[B] 12   420 ms   579 ms   531 ms  197-84-208-146.cpt.mweb.co.za [197.84.208.146]
 13   494 ms   567 ms   540 ms  41-86-110-200.mweb.co.za [41.86.110.200][[/B]

Trace to bbc.co.uk:

Code: 
Tracing route to bbc.co.uk [151.101.192.81]
over a maximum of 30 hops:

  1    <1 ms    <1 ms    <1 ms  router.asus.com [192.168.10.1]
  2     *        *        *     Request timed out.
  3     8 ms     6 ms     5 ms  196.38.75.121
  4    58 ms    63 ms    66 ms  196.38.75.122
  5    22 ms    20 ms    20 ms  196.38.77.88
  6     5 ms     5 ms     5 ms  mi-za-cpt-p7-te0-7-0-0.ip.isnet.net [168.209.2.12]
  7     5 ms     5 ms     6 ms  168.209.2.130
[B]  8   507 ms   368 ms   319 ms  196-10-140-103.ixp.capetown [196.10.140.103][/B]
  9     3 ms     5 ms     7 ms  192.168.100.6
[B] 10   322 ms   367 ms   455 ms  151.101.192.81[/B]
 
Last edited:
Joker said:
Trace to www.news24.com:

Code: 
Tracing route to www.news24.com [41.86.110.200]
over a maximum of 30 hops:

  1     1 ms     1 ms     1 ms  router.asus.com [192.168.10.1]
  2     *        *        *     Request timed out.
  3     4 ms     5 ms     6 ms  196.38.75.121
  4     *       20 ms    21 ms  196.38.75.122
  5    23 ms    27 ms    25 ms  196.38.77.88
  6     5 ms     5 ms     5 ms  mi-za-cpt-p8-te0-0-0-0.ip.isnet.net [168.209.6.13]
[B]  7   419 ms   475 ms   665 ms  168.209.2.131[/B]
  8     6 ms     4 ms     5 ms  mweb-2.cinx.net.za [196.223.22.25]
  9     7 ms     7 ms     5 ms  197-84-7-54.cpt.mweb.co.za [197.84.7.54]
 10    13 ms    11 ms    10 ms  197-84-247-89.cpt.mweb.co.za [197.84.247.89]
 11     7 ms     6 ms     4 ms  OPTI-CPT-N1-Hosting.optinet.net [197.84.196.6]
[B] 12   420 ms   579 ms   531 ms  197-84-208-146.cpt.mweb.co.za [197.84.208.146]
 13   494 ms   567 ms   540 ms  41-86-110-200.mweb.co.za [41.86.110.200][[/B]
Click to expand...

hop 6 == peering router via cinx to optinet's network
hop 12/13 == optinet's network towards their client news24

let me chat to the optinet guys and find out the reason for the latency.
 
Last edited:
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X