Status
Not open for further replies.
So was Openweb lol

Yeah but OW was run by douches who also didn't know how to do anything other than resell accounts. CW is DJ's baby. Before he was banned he came across as incredibly passionate, hardworking and knowledgeable about the ISP business. It just doesn't make sense.

Thank god for Cool Ideas I guess. I was not keen to go to Afrihost, Axxess, Vox etc.
 
Ping went to 150ms to CS GO servers earlier and is still around 125ms :(

Speedtest :erm:

Capture.JPG

Going to reboot everything.
 
today has been terrible for streaming from international sites
 
A traceroute to Twitch looks bad:
Code:
mithrandi@lorien ~> mtr -T --report 52.223.194.76          
Start: Sun Oct  1 20:55:02 2017                            
HOST: lorien                      Loss%   Snt   Last   Avg  Best  Wrst StDev                                           
  1.|-- elvandar.mithrandi.net     0.0%    10    0.6   0.5   0.3   0.6   0.0                                           
  2.|-- 154.0.4.1                  0.0%    10   72.3  72.7  60.3  85.0   6.7                                           
  3.|-- 154.0.2.1                  0.0%    10  123.3 120.5 102.0 146.8  14.6                                           
  4.|-- 154.0.2.125                0.0%    10  147.4 132.9 109.5 147.4  13.5                                           
  5.|-- 154.0.1.66                 0.0%    10  122.9 138.2 122.9 158.2  11.5                                           
  6.|-- 154.0.1.142               20.0%    10  144.6 1901. 104.9 7248. 2534.9                                          
  7.|-- 154.0.1.146                0.0%    10  284.9 299.0 271.8 318.5  16.7                                           
  8.|-- lhr01-br01.justin.tv       0.0%    10  305.9 299.8 264.5 332.8  21.6                                           
  9.|-- 10.26.52.138               0.0%    10  278.6 297.5 260.5 326.3  20.8
 10.|-- video-edge-c67a70.lhr03.h  0.0%    10  401.6 368.8 276.1 401.7  53.0

1. First hop (after my router) latency is pretty bad (this is all cabled).

2. [-]The latency jumps up by ~300ms later on which smells like traffic traversing the international link more than once.[/-] Nevermind, the first two hops' latency is too inconsistent to draw conclusions about the rest.
 
Last edited:
Post reboot of CPE, router and switch, latency is still the same - bandwidth is fine. Weird.

Ticket Created #COOL-20171001-40830
 
Looks to be a Vumatel related issue. We have logged it with them.
 
Also seeing a huge jump in latency to Telkom

Code:
                                                        Packets               Pings
 Host                                                 Loss%   Snt   Last   Avg  Best  Wrst StDev
 1. 155.93.141.1                                       0.0%    28    0.9   2.9   0.9  12.7   2.1
 2. c4x-backbone.coolideas.co.za                       0.0%    28    2.5   2.6   1.1   7.1   1.3
 3. cx-backbone.coolideas.co.za                        0.0%    28    2.9   4.1   1.1  35.5   6.2
 4. c3u-backbone.coolideas.co.za                       0.0%    28    2.9   2.9   1.2   5.0   0.8
 5. 197.96.214.112                                     0.0%    27    2.7   3.6   2.3   7.0   1.1
 6. 196.26.0.130                                       0.0%    27    3.9   3.8   2.4  12.0   1.9
 7. openserve.ixp.joburg                               3.7%    27  168.3 167.3 166.0 171.6   1.1
 8. rrba-ip-lir-1-gig-0-0-0-102.telkom-ipnet.co.za     0.0%    27  164.6 165.2 164.4 167.6   0.8
 9. nbsc-ip-er-1-pos-8-0-0.telkom-ipnet.co.za          0.0%    27  168.9 170.6 168.8 175.9   1.5
10. 196.25.10.42                                       0.0%    27  169.1 169.2 167.2 180.0   2.8
11. 198.54.206.73                                      0.0%    27  167.6 168.5 167.1 171.5   1.2
12. www.telkom.co.za                                   0.0%    27  170.9 173.2 169.3 224.4  10.4

Something is definitely not right
 
Yeah local latency is like 160ms and international is like minutes, not ms.... In my experience on the 200mbps package and Vumatel, almost every Sunday evening I have been having crap speeds, issues with latencies etc... I see the same pattern as ADSL on telkom... Contact Cool Ideas they say its Vumatels issue, call Vumatel they say its Cool Ideas Issue... I had enough of it... I think Cool Ideas and Vumatel are just expanding so fast that their infrastructure is not capable of giving a good stable performance all the time, as its always some issue.
 
A traceroute to Twitch looks bad:
Code:
mithrandi@lorien ~> mtr -T --report 52.223.194.76          
Start: Sun Oct  1 20:55:02 2017                            
HOST: lorien                      Loss%   Snt   Last   Avg  Best  Wrst StDev                                           
  1.|-- elvandar.mithrandi.net     0.0%    10    0.6   0.5   0.3   0.6   0.0                                           
  2.|-- 154.0.4.1                  0.0%    10   72.3  72.7  60.3  85.0   6.7                                           
  3.|-- 154.0.2.1                  0.0%    10  123.3 120.5 102.0 146.8  14.6                                           
  4.|-- 154.0.2.125                0.0%    10  147.4 132.9 109.5 147.4  13.5                                           
  5.|-- 154.0.1.66                 0.0%    10  122.9 138.2 122.9 158.2  11.5                                           
  6.|-- 154.0.1.142               20.0%    10  144.6 1901. 104.9 7248. 2534.9                                          
  7.|-- 154.0.1.146                0.0%    10  284.9 299.0 271.8 318.5  16.7                                           
  8.|-- lhr01-br01.justin.tv       0.0%    10  305.9 299.8 264.5 332.8  21.6                                           
  9.|-- 10.26.52.138               0.0%    10  278.6 297.5 260.5 326.3  20.8
 10.|-- video-edge-c67a70.lhr03.h  0.0%    10  401.6 368.8 276.1 401.7  53.0

1. First hop (after my router) latency is pretty bad (this is all cabled).

2. [-]The latency jumps up by ~300ms later on which smells like traffic traversing the international link more than once.[/-] Nevermind, the first two hops' latency is too inconsistent to draw conclusions about the rest.

Definitely looks like Vumatel issue on your side. Look at how it looks in CTN:
Code:
~ $ mtr -T --report-wide --show-ips 52.223.194.76         
Start: Sun Oct  1 21:30:34 2017
HOST: outlawstar                                                Loss%   Snt   Last   Avg  Best  Wrst StDev
  1.|-- router.heaven.za.net (192.168.241.1)                       0.0%    10    0.4   0.3   0.3   0.4   0.0
  2.|-- 155.93.252.1                                               0.0%    10    1.3   1.3   1.2   1.4   0.0
  3.|-- c3h-backbone.coolideas.co.za (154.0.1.125)                 0.0%    10    1.1   1.1   0.9   1.2   0.0
  4.|-- cd-backbone.coolideas.co.za (154.0.1.13)                   0.0%    10    1.3   1.3   1.2   1.4   0.0
  5.|-- wc-grwf-ar2-eth1-1701.enetworks.net (45.220.191.17)        0.0%    10    1.4   1.4   1.3   1.5   0.0
  6.|-- wc-grwf-br1-gi0-2-3202.enetworks.net (154.72.97.164)       0.0%    10    1.7   1.8   1.5   2.0   0.0
  7.|-- 41.60.242.1                                                0.0%    10    2.6   2.3   2.0   3.3   0.3
  8.|-- teng0-3-0-5.luk-p1-tho.liquidtelecom.net (5.11.10.170)     0.0%    10  142.2 142.1 141.8 142.9   0.0
  9.|-- bundle-ether3-luk-pe1-tho.liquidtelecom.net (5.11.10.103)  0.0%    10  142.2 142.1 141.9 142.3   0.0
 10.|-- ten0-0-2-0-luk-pr1-tho.liquidtelecom.net (5.11.10.202)     0.0%    10  141.9 142.1 141.7 142.5   0.0
 11.|-- lhr01-br01.justin.tv (195.66.225.174)                      0.0%    10  142.2 142.3 142.0 142.5   0.0
 12.|-- ???                                                       100.0    10    0.0   0.0   0.0   0.0   0.0
 13.|-- video-edge-c67a70.lhr03.hls.ttvnw.net (52.223.194.76)      0.0%    10  154.4 159.0 154.3 200.6  14.6
 
Also seeing a huge jump in latency to Telkom

Code:
                                                        Packets               Pings
 Host                                                 Loss%   Snt   Last   Avg  Best  Wrst StDev
 1. 155.93.141.1                                       0.0%    28    0.9   2.9   0.9  12.7   2.1
 2. c4x-backbone.coolideas.co.za                       0.0%    28    2.5   2.6   1.1   7.1   1.3
 3. cx-backbone.coolideas.co.za                        0.0%    28    2.9   4.1   1.1  35.5   6.2
 4. c3u-backbone.coolideas.co.za                       0.0%    28    2.9   2.9   1.2   5.0   0.8
 5. 197.96.214.112                                     0.0%    27    2.7   3.6   2.3   7.0   1.1
 6. 196.26.0.130                                       0.0%    27    3.9   3.8   2.4  12.0   1.9
 7. openserve.ixp.joburg                               3.7%    27  168.3 167.3 166.0 171.6   1.1
 8. rrba-ip-lir-1-gig-0-0-0-102.telkom-ipnet.co.za     0.0%    27  164.6 165.2 164.4 167.6   0.8
 9. nbsc-ip-er-1-pos-8-0-0.telkom-ipnet.co.za          0.0%    27  168.9 170.6 168.8 175.9   1.5
10. 196.25.10.42                                       0.0%    27  169.1 169.2 167.2 180.0   2.8
11. 198.54.206.73                                      0.0%    27  167.6 168.5 167.1 171.5   1.2
12. www.telkom.co.za                                   0.0%    27  170.9 173.2 169.3 224.4  10.4

Something is definitely not right

This is because CISP uses HurricaneElectric as their international transit provider, who has a POP at NapAfrica JHB, where SAIX/OpenServe peer with HurricaneElectric.

See the traceroute back:
Code:
tpr-route-server>traceroute 155.93.141.1

Type escape sequence to abort.
Tracing the route to 155.93.141.1

  1 tpr-ip-esr-1-fe-1-0-3.telkom-ipnet.co.za (196.25.9.45) 0 msec 0 msec 0 msec
  2 rrba-ip-se-2-gig-2-0-0-102.telkom-ipnet.co.za (196.43.25.138) 0 msec 4 msec 0 msec
  3 he.ixp.joburg (196.60.8.42) 4 msec 4 msec 0 msec
  4 10ge3-11.core1.lon2.he.net (184.105.81.253) [AS 6939] 176 msec 204 msec 200 msec
  5 10ge8-4.core1.lon1.he.net (72.52.92.221) [AS 6939] 164 msec 164 msec 164 msec
  6 cool-ideas-service-provider.10gigabitethernet5-4.core1.lon1.he.net (216.66.80.170) [AS 6939] 168 msec 164 msec 164 msec
  7 c3y-backbone.coolideas.co.za (154.0.1.142) [AS 37680] 168 msec 164 msec 168 msec
  8 c1u-backbone.coolideas.co.za (154.0.1.66) [AS 37680] 164 msec 164 msec 172 msec
  9 cy-backbone.coolideas.co.za (154.0.2.34) [AS 37680] 168 msec 176 msec 168 msec
 10  *  *  *

Looks like HurricaneElectric is a bad choice as international transit provider. Personally I am a fan of Telia and Level3. Level3 unfortunately also has pops in SA.

The Telkom/SAIX/OpenServe peering issue has been going on for a while now. Would be awesome if CISP can eventually resolve it.
 
Most definitely a Vumatel issue, other networks aren't affected. Vumatel trenched specifically. It has been logged with the senior guys in Sweden that manage the network. This happened before they did the logical network split previously.
 
Yeah local latency is like 160ms and international is like minutes, not ms.... In my experience on the 200mbps package and Vumatel, almost every Sunday evening I have been having crap speeds, issues with latencies etc... I see the same pattern as ADSL on telkom... Contact Cool Ideas they say its Vumatels issue, call Vumatel they say its Cool Ideas Issue... I had enough of it... I think Cool Ideas and Vumatel are just expanding so fast that their infrastructure is not capable of giving a good stable performance all the time, as its always some issue.

last sunday i watched nfl quite perfectly
 
Most definitely a Vumatel issue, other networks aren't affected. Vumatel trenched specifically. It has been logged with the senior guys in Sweden that manage the network. This happened before they did the logical network split previously.

that's still **** for us
 
Most definitely a Vumatel issue, other networks aren't affected. Vumatel trenched specifically. It has been logged with the senior guys in Sweden that manage the network. This happened before they did the logical network split previously.

Same for Openserve. 400ms to EU
 
Capture.PNG
Vumatel AE, Not CISP but another IS-based ISP.

From my traceroutes, it was an issue inside IS.
1 <1 ms <1 ms <1 ms 192.168.1.1
2 <1 ms <1 ms <1 ms xxxx.dyn.ftth.broadband.is [xxxx]
3 63 ms 32 ms 30 ms 197.97.90.33
4 97 ms 96 ms 104 ms cdsl2-rba-vl151.ip.isnet.net [196.38.73.10]
5 90 ms 95 ms 97 ms 196.38.73.157
6 105 ms 103 ms 105 ms core2b-pkl-te0-0-0-0.ip.isnet.net [196.26.0.63]


Started 19:30, ended 21:30
 
Last edited:
View attachment 466907
Vumatel AE, Not CISP but another IS-based ISP.

From my traceroutes, it was an issue inside IS.
1 <1 ms <1 ms <1 ms 192.168.1.1
2 <1 ms <1 ms <1 ms xxxx.dyn.ftth.broadband.is [xxxx]
3 63 ms 32 ms 30 ms 197.97.90.33
4 97 ms 96 ms 104 ms cdsl2-rba-vl151.ip.isnet.net [196.38.73.10]
5 90 ms 95 ms 97 ms 196.38.73.157
6 105 ms 103 ms 105 ms core2b-pkl-te0-0-0-0.ip.isnet.net [196.26.0.63]


Started 19:30, ended 21:30
Nah would still be Vumatel, our issue isn't IS related.
 
This is because CISP uses HurricaneElectric as their international transit provider, who has a POP at NapAfrica JHB, where SAIX/OpenServe peer with HurricaneElectric.

See the traceroute back:
Code:
tpr-route-server>traceroute 155.93.141.1

Type escape sequence to abort.
Tracing the route to 155.93.141.1

  1 tpr-ip-esr-1-fe-1-0-3.telkom-ipnet.co.za (196.25.9.45) 0 msec 0 msec 0 msec
  2 rrba-ip-se-2-gig-2-0-0-102.telkom-ipnet.co.za (196.43.25.138) 0 msec 4 msec 0 msec
  3 he.ixp.joburg (196.60.8.42) 4 msec 4 msec 0 msec
  4 10ge3-11.core1.lon2.he.net (184.105.81.253) [AS 6939] 176 msec 204 msec 200 msec
  5 10ge8-4.core1.lon1.he.net (72.52.92.221) [AS 6939] 164 msec 164 msec 164 msec
  6 cool-ideas-service-provider.10gigabitethernet5-4.core1.lon1.he.net (216.66.80.170) [AS 6939] 168 msec 164 msec 164 msec
  7 c3y-backbone.coolideas.co.za (154.0.1.142) [AS 37680] 168 msec 164 msec 168 msec
  8 c1u-backbone.coolideas.co.za (154.0.1.66) [AS 37680] 164 msec 164 msec 172 msec
  9 cy-backbone.coolideas.co.za (154.0.2.34) [AS 37680] 168 msec 176 msec 168 msec
 10  *  *  *

Looks like HurricaneElectric is a bad choice as international transit provider. Personally I am a fan of Telia and Level3. Level3 unfortunately also has pops in SA.

The Telkom/SAIX/OpenServe peering issue has been going on for a while now. Would be awesome if CISP can eventually resolve it.
Which Telkom/Openserve issue. Can you PM me an example. It's nothing to do with HE but more often the ISP peering with HE follows their peering guidelines which sets a pref incorrectly.
 
Status
Not open for further replies.
Top
Sign up to the MyBroadband newsletter
X