Afrihost - Pure Fibre Feedback Thread

Status
Not open for further replies.
Loppas said:
If you are referring to cape town routing, then yes it does go direct to london. You can select between their different nodes at the top (cpt, joburg or durban), by default its set to joburg.
Click to expand...

I didn't even notice that :laugh:

Yeah, their latency from Cape Town is not only lower, but stable. So cheaper and more optimised routing. I'm not inclined to risk a migration at the end of the year, but this is food for thought.
 
3HH said:
@AfriMan, I don't really want to jump on the bandwagon, I saw your post from a few days ago where you mentioned that you guys are still tweaking things, so I'm inclined to wait a little while longer to see if things improve. However, please understand the client's point of view as well, that if we need to pay extra each month for a VPN just to get our latency down to a more usable level, then why not just move to a slightly more expensive ISP that already does the routing for us, making it one less thing we need to worry about?

For interest sake, here are my results when I ping the latest IP in the thread -

Pinging 139.162.220.124 with 32 bytes of data:
Reply from 139.162.220.124: bytes=32 time=192ms TTL=53
Reply from 139.162.220.124: bytes=32 time=201ms TTL=53
Reply from 139.162.220.124: bytes=32 time=172ms TTL=53
Reply from 139.162.220.124: bytes=32 time=185ms TTL=53

Ping statistics for 139.162.220.124:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 172ms, Maximum = 201ms, Average = 187ms
Click to expand...

I totally get that. I know our NOC and network managers do listen to these convos on MyBB and they are always tweaking for best performance.
 
Ecco said:
Hey guys, last night my router starts randomly rebooting. while watching 1 episode on netflix, the router did a reboot 3 times. At which point i gave up and and went to bed. Anybody else experienced similar? Router is the D-Link DIR-825
Click to expand...
I reported the above a while back and this is now becoming more frequent. Last night router rebooted itself twice in one hour.
Previous night about 3 times during 2 hours. I updated the firmware when I received the router.

Any suggestions @AfriMan ?
 
Latency back up to 176ms when it was 143ms a few hours ago. It's not even evening time yet. If this is the new norm, then however much I've been a fan of the rest of AH's network performance, this is a deal-breaker for me.
 
Seeyou said:
Latency back up to 176ms when it was 143ms a few hours ago. It's not even evening time yet. If this is the new norm, then however much I've been a fan of the rest of AH's network performance, this is a deal-breaker for me.
Click to expand...
yip, been using vpn the whole day.
 
The thing is congestion by the looks of it.

Afrihost
Axxess
Webafrica
Telkom ISP
Telkom Mobile

All of these ISPs uses the same Openserve transit and by same transit it's the same routers too.

Have a look here @blunt tracert from CPT to riotgames

please help check latency and route for different ISP's

Hi All, If at all possible i would like to gather some information around routing and latency to - > 185.40.64.65 If you can do a normal ping and tracert and post the results with your ISP that would be great. This is to -> prod.euw1.lol.riotgames.com I am trying to decide on a new fibre ISP...
mybroadband.co.za mybroadband.co.za

Look @Geoff.D on Telkom Fibre in PTA East

please help check latency and route for different ISP's

Hi All, If at all possible i would like to gather some information around routing and latency to - > 185.40.64.65 If you can do a normal ping and tracert and post the results with your ISP that would be great. This is to -> prod.euw1.lol.riotgames.com I am trying to decide on a new fibre ISP...
mybroadband.co.za mybroadband.co.za

Look at my trace route on Afrihost in PTA
758626



All these ISPs and in all regions seem to be going through 10.189.30.10 So unless that internal ip is actually different routers or not I don't know but yeah that is a lot of ISPs and a lot of traffic going though there.

Honestly I don't think Openserve transit was the best decision. There are better options out there. Network Platforms, Cogent, HE, and even WIOCC which Rain is using with some very decent routing. I can't believe for one second that Openserve offered the better deal be it price or routing to Afrihost.
 
cavedog said:
The thing is congestion by the looks of it.

Afrihost
Axxess
Webafrica
Telkom ISP
Telkom Mobile

All of these ISPs uses the same Openserve transit and by same transit it's the same routers too.
-SNIP-
Click to expand...

I typed up a long response about how there is no way they are using OpenServe transit. But it actually checks out.

Mine, Afrihost, JHB:
1 102-182-102-1.ip.afrihost.co.za (102.182.102.1) 19.668 ms 3.161 ms 3.153 ms
2 tobitropin.net.afrihost.co.za (169.1.1.161) 3.555 ms 2.649 ms 2.085 ms
3 lucopel.net.afrihost.co.za (169.1.1.129) 1.212 ms 1.311 ms 1.166 ms
4 169-1-21-101.ip.afrihost.co.za (169.1.21.101) 2.395 ms 2.198 ms 3.334 ms
** 5 rndf-ip-ll-1-wan.telkom-ipnet.co.za (196.25.199.141) 2.697 ms 5.659 ms 3.537 ms
6 10.189.30.10 (10.189.30.10) 185.838 ms 181.912 ms 182.023 ms
7 * * *
8 * * *
9 * * *
10 104.160.141.103 (104.160.141.103) 185.548 ms 190.706 ms 189.108 ms
11 104.160.141.107 (104.160.141.107) 190.359 ms 186.049 ms 185.766 ms
12 185.40.64.65 (185.40.64.65) 189.236 ms 190.173 ms 189.770 ms

I seriously doubt that they are doing transit with Openserve. Almost 100% promise that it wont be the case.
But OpenServe likely advertise it as a route, and Afrihost learnt it. Likely, unintentionally?
@AfriMan, any ideas here?
 
Echo, which is used by Afrihost, Axxess and Web Africa all use SAIX for transit. SAIX is actually a fantastic network, in addition to Echo's massive peering at the IX is a great thing for customers.
 
Ecco said:
I reported the above a while back and this is now becoming more frequent. Last night router rebooted itself twice in one hour.
Previous night about 3 times during 2 hours. I updated the firmware when I received the router.

Any suggestions @AfriMan ?
Click to expand...

I think the best place to start would be D-Link support. Our warehouse manager says they are very keen to get directly involved with issues like this - and they can send you patches directly if that fixes the issue. If they can't, then we'll arrange the return/exchange on their behalf.

www.d-link.co.za

Support - D-Link Southern Africa

Contact Us Business Hours: Mon - Fri 08:30 - 16:30 Afterhour Support: Mon - Fri 17:00 - 20:30 Sat - Sun 09:00 - 18:00 08600 35465 (DLINK) or 012 741 2000 Email Us Sales: [email protected] Support: [email protected]: [email protected]
www.d-link.co.za www.d-link.co.za
 
brentm said:
I typed up a long response about how there is no way they are using OpenServe transit. But it actually checks out.

Mine, Afrihost, JHB:
1 102-182-102-1.ip.afrihost.co.za (102.182.102.1) 19.668 ms 3.161 ms 3.153 ms
2 tobitropin.net.afrihost.co.za (169.1.1.161) 3.555 ms 2.649 ms 2.085 ms
3 lucopel.net.afrihost.co.za (169.1.1.129) 1.212 ms 1.311 ms 1.166 ms
4 169-1-21-101.ip.afrihost.co.za (169.1.21.101) 2.395 ms 2.198 ms 3.334 ms
** 5 rndf-ip-ll-1-wan.telkom-ipnet.co.za (196.25.199.141) 2.697 ms 5.659 ms 3.537 ms
6 10.189.30.10 (10.189.30.10) 185.838 ms 181.912 ms 182.023 ms
7 * * *
8 * * *
9 * * *
10 104.160.141.103 (104.160.141.103) 185.548 ms 190.706 ms 189.108 ms
11 104.160.141.107 (104.160.141.107) 190.359 ms 186.049 ms 185.766 ms
12 185.40.64.65 (185.40.64.65) 189.236 ms 190.173 ms 189.770 ms

I seriously doubt that they are doing transit with Openserve. Almost 100% promise that it wont be the case.
But OpenServe likely advertise it as a route, and Afrihost learnt it. Likely, unintentionally?
@AfriMan, any ideas here?
Click to expand...

I believe we have moved our international routing from Liquid Telecom to Telkom, but our local routing is as it always was before.
 
@AfriMan Its been brought up already, but is the network team looking into why the latency rises by 20ms in the evening and it appears to be all over the place too.

E.g.
Code: 
Pinging comedy.co.uk [91.192.195.29] with 32 bytes of data:
Reply from 91.192.195.29: bytes=32 time=155ms TTL=55
Reply from 91.192.195.29: bytes=32 time=166ms TTL=55
Reply from 91.192.195.29: bytes=32 time=164ms TTL=55
Reply from 91.192.195.29: bytes=32 time=167ms TTL=55
Reply from 91.192.195.29: bytes=32 time=171ms TTL=55
Reply from 91.192.195.29: bytes=32 time=177ms TTL=55
Reply from 91.192.195.29: bytes=32 time=165ms TTL=55
Reply from 91.192.195.29: bytes=32 time=158ms TTL=55
Reply from 91.192.195.29: bytes=32 time=172ms TTL=55
Reply from 91.192.195.29: bytes=32 time=161ms TTL=55
Reply from 91.192.195.29: bytes=32 time=174ms TTL=55
Reply from 91.192.195.29: bytes=32 time=201ms TTL=55
Reply from 91.192.195.29: bytes=32 time=190ms TTL=55
Reply from 91.192.195.29: bytes=32 time=174ms TTL=55
Reply from 91.192.195.29: bytes=32 time=209ms TTL=55
Reply from 91.192.195.29: bytes=32 time=162ms TTL=55
Reply from 91.192.195.29: bytes=32 time=172ms TTL=55
Reply from 91.192.195.29: bytes=32 time=197ms TTL=55
Reply from 91.192.195.29: bytes=32 time=173ms TTL=55
Reply from 91.192.195.29: bytes=32 time=156ms TTL=55
Reply from 91.192.195.29: bytes=32 time=171ms TTL=55
Reply from 91.192.195.29: bytes=32 time=167ms TTL=55
Reply from 91.192.195.29: bytes=32 time=166ms TTL=55
Reply from 91.192.195.29: bytes=32 time=166ms TTL=55
Reply from 91.192.195.29: bytes=32 time=171ms TTL=55
Reply from 91.192.195.29: bytes=32 time=173ms TTL=55
Reply from 91.192.195.29: bytes=32 time=174ms TTL=55
Reply from 91.192.195.29: bytes=32 time=181ms TTL=55
Reply from 91.192.195.29: bytes=32 time=176ms TTL=55
Reply from 91.192.195.29: bytes=32 time=182ms TTL=55
Reply from 91.192.195.29: bytes=32 time=158ms TTL=55

Ping statistics for 91.192.195.29:
    Packets: Sent = 31, Received = 31, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 155ms, Maximum = 209ms, Average = 172ms

This is what im getting atm whereas it would usually be 148ms~. The ping should not be so sporadic either.
 
Loppas said:
@AfriMan Its been brought up already, but is the network team looking into why the latency rises by 20ms in the evening and it appears to be all over the place too.

E.g.
Code: 
Pinging comedy.co.uk [91.192.195.29] with 32 bytes of data:
Reply from 91.192.195.29: bytes=32 time=155ms TTL=55
Reply from 91.192.195.29: bytes=32 time=166ms TTL=55
Reply from 91.192.195.29: bytes=32 time=164ms TTL=55
Reply from 91.192.195.29: bytes=32 time=167ms TTL=55
Reply from 91.192.195.29: bytes=32 time=171ms TTL=55
Reply from 91.192.195.29: bytes=32 time=177ms TTL=55
Reply from 91.192.195.29: bytes=32 time=165ms TTL=55
Reply from 91.192.195.29: bytes=32 time=158ms TTL=55
Reply from 91.192.195.29: bytes=32 time=172ms TTL=55
Reply from 91.192.195.29: bytes=32 time=161ms TTL=55
Reply from 91.192.195.29: bytes=32 time=174ms TTL=55
Reply from 91.192.195.29: bytes=32 time=201ms TTL=55
Reply from 91.192.195.29: bytes=32 time=190ms TTL=55
Reply from 91.192.195.29: bytes=32 time=174ms TTL=55
Reply from 91.192.195.29: bytes=32 time=209ms TTL=55
Reply from 91.192.195.29: bytes=32 time=162ms TTL=55
Reply from 91.192.195.29: bytes=32 time=172ms TTL=55
Reply from 91.192.195.29: bytes=32 time=197ms TTL=55
Reply from 91.192.195.29: bytes=32 time=173ms TTL=55
Reply from 91.192.195.29: bytes=32 time=156ms TTL=55
Reply from 91.192.195.29: bytes=32 time=171ms TTL=55
Reply from 91.192.195.29: bytes=32 time=167ms TTL=55
Reply from 91.192.195.29: bytes=32 time=166ms TTL=55
Reply from 91.192.195.29: bytes=32 time=166ms TTL=55
Reply from 91.192.195.29: bytes=32 time=171ms TTL=55
Reply from 91.192.195.29: bytes=32 time=173ms TTL=55
Reply from 91.192.195.29: bytes=32 time=174ms TTL=55
Reply from 91.192.195.29: bytes=32 time=181ms TTL=55
Reply from 91.192.195.29: bytes=32 time=176ms TTL=55
Reply from 91.192.195.29: bytes=32 time=182ms TTL=55
Reply from 91.192.195.29: bytes=32 time=158ms TTL=55

Ping statistics for 91.192.195.29:
    Packets: Sent = 31, Received = 31, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 155ms, Maximum = 209ms, Average = 172ms

This is what im getting atm whereas it would usually be 148ms~. The ping should not be so sporadic either.
Click to expand...

Run a trace to bras.afrihost.com and to 1.1.1.1 at the same time and post them here pls
 
waylander said:
Run a trace to bras.afrihost.com and to 1.1.1.1 at the same time and post them here pls
Click to expand...

I cant do winmtr to both at the same time without the router timing out on one of the windows, so i did this:

Code: 
|------------------------------------------------------------------------------------------|
|                                      WinMTR statistics                                   |
|                       Host              -   %  | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
|                             192.168.0.1 -    0 |   75 |   75 |    0 |    0 |    9 |    0 |
|                   No response from host -  100 |   16 |    0 |    0 |    0 |    0 |    0 |
|         165-73-14-113.ip.afrihost.co.za -    0 |   75 |   75 |    3 |    4 |   40 |    4 |
|            169-1-21-2.ip.afrihost.co.za -    0 |   75 |   75 |    2 |    2 |    6 |    2 |
|            169-1-21-3.ip.afrihost.co.za -    0 |   75 |   75 |    2 |    2 |    4 |    3 |
|                 cloudflare.ixp.capetown -    0 |   75 |   75 |    2 |    2 |    4 |    3 |
|                         one.one.one.one -    0 |   75 |   75 |    2 |    2 |    3 |    2 |
|________________________________________________|______|______|______|______|______|______|

Ping statistics for 155.239.255.250:
    Packets: Sent = 77, Received = 77, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 1ms, Maximum = 11ms, Average = 2ms


Its not congestion on my side, the root of the issues starts once it hits international:
Code: 
|------------------------------------------------------------------------------------------|
|                                      WinMTR statistics                                   |
|                       Host              -   %  | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
|                             192.168.0.1 -    0 |   41 |   41 |    0 |    0 |    1 |    0 |
|                   No response from host -  100 |    9 |    0 |    0 |    0 |    0 |    0 |
|         165-73-14-113.ip.afrihost.co.za -    0 |   41 |   41 |    3 |    7 |   39 |   31 |
|            169-1-21-2.ip.afrihost.co.za -    0 |   41 |   41 |    2 |    2 |    4 |    2 |
|            169-1-21-3.ip.afrihost.co.za -    0 |   41 |   41 |    2 |    2 |    4 |    3 |
|                            196.25.26.65 -    0 |   41 |   41 |    3 |    6 |   32 |    3 |
|                             10.189.30.2 -    0 |   41 |   41 |  143 |  154 |  190 |  158 |
|                          195.66.227.119 -    0 |   41 |   41 |  144 |  152 |  183 |  151 |
|                 te4-4-rt01-dcn1.cwcs.uk -    0 |   41 |   41 |  147 |  171 |  341 |  148 |
|                           comedy.org.uk -    0 |   41 |   41 |  148 |  164 |  190 |  156 |
|________________________________________________|______|______|______|______|______|______|
 
waylander said:
Echo, which is used by Afrihost, Axxess and Web Africa all use SAIX for transit. SAIX is actually a fantastic network, in addition to Echo's massive peering at the IX is a great thing for customers.
Click to expand...

What makes Telkom/SAIX transit so good? I mean Telkom is at LINX but hands off traffic to PCCW in London to a destination in London when just a week ago they peered directly unless maybe coreix stopped that for whatever reason.

There are so so so many transit providers so not sure why Telkom is better.
 
cavedog said:
What makes Telkom/SAIX transit so good? I mean Telkom is at LINX but hands off traffic to PCCW in London to a destination in London when just a week ago they peered directly unless maybe coreix stopped that for whatever reason.

There are so so so many transit providers so not sure why Telkom is better.
Click to expand...

I'm not familiar with your routing to CoreIX, this is mine:

1 * * *
2 165-73-14-17.ip.afrihost.co.za (165.73.14.17) 4.419 ms 3.785 ms 3.947 ms
3 169-1-21-234.ip.afrihost.co.za (169.1.21.234) 3.599 ms 3.159 ms 3.181 ms
4 169-1-21-235.ip.afrihost.co.za (169.1.21.235) 3.348 ms 3.442 ms 3.147 ms
5 rndf-ip-ll-1-wan.telkom-ipnet.co.za (196.25.199.141) 3.477 ms 4.219 ms 4.102 ms
6 10.189.30.2 (10.189.30.2) 196.590 ms 199.993 ms 201.054 ms
7 ge-0-1-0-68.peering1.the.lon1.coreix.net (195.66.225.58) 198.111 ms 199.997 ms 204.624 ms
8 xe-5-1.edge4.enf.lon5.coreix.net (89.187.93.9) 201.454 ms 185.689 ms 185.819 ms
9 85.13.221.178.reverse.coreix.net (85.13.221.178) 188.024 ms 194.839 ms 196.580 ms
Click to expand...

Their upstream diversity (NTT, who owns IS, Cogent, BTN, L3 and IS) with POP in AMX, LON, CIX, FR and more is fantastic.
 
@AfriMan just as a matter of interest - random question - my parents changed to a Tenda router as the WiFi is *much* better than the DIR-825 (side note), when I activated the new router (Frogfoot) with AH they were given an IP 165.73.120.xxx which GeoIP looks up as Western Cape, previously they had 169.1.253.xxx which GeoIP looks up as JHB - my uncle who is also on FF about 5 mins away is still on his DIR-825 and has one of those JHB IP's.

I'm on Openserve and have a 169.0.116.xxx IP which geoip's as CPT.

Is this just luck of the draw, or is there a reason why the Afrihost DIR's both got the JHB IP's and this new router got a CPT one on a different range?
 
blunt said:
@AfriMan just as a matter of interest - random question - my parents changed to a Tenda router as the WiFi is *much* better than the DIR-825 (side note), when I activated the new router (Frogfoot) with AH they were given an IP 165.73.120.xxx which GeoIP looks up as Western Cape, previously they had 169.1.253.xxx which GeoIP looks up as JHB - my uncle who is also on FF about 5 mins away is still on his DIR-825 and has one of those JHB IP's.

I'm on Openserve and have a 169.0.116.xxx IP which geoip's as CPT.

Is this just luck of the draw, or is there a reason why the Afrihost DIR's both got the JHB IP's and this new router got a CPT one on a different range?
Click to expand...

IP geo-location just depends on the prefix you're assigned and shouldn't have any impact on your actual experience.
 
Status
Not open for further replies.
Back
Top
Sign up to the MyBroadband newsletter
X