Cool Ideas Fibre ISP – Feedback Thread 4

razed

Well-Known Member
Joined
Dec 3, 2008
Messages
129
11521841987.png
 

sn0rbaard

Active Member
Joined
Aug 9, 2017
Messages
40
i3d peer locally so typically this means their local cluster went offline and they routed via international.
@PBCool I mean no disrespect but it does not make sense, please explain? It is the fourth day that I get "international" 170ms instead of 20ms ping.

All my friends on competing local ISPs has no such issue. And on our Rocket League group I got no confirmation from anyone else having this issue. Please have a look at the following two WinMTRs.

Here is tonight's WinMTR via my Cool Ideas account (versus competing ISP after):

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| 192.168.1.1 - 0 | 129 | 129 | 0 | 0 | 1 | 0 |
| u6q-cust.coolideas.co.za - 0 | 129 | 129 | 1 | 2 | 37 | 2 |
| 100.98.0.1 - 0 | 129 | 129 | 1 | 2 | 38 | 2 |
| 100.98.0.49 - 0 | 129 | 129 | 16 | 17 | 52 | 18 |
| 100.99.0.180 - 0 | 129 | 129 | 16 | 17 | 51 | 18 |
| i3d.jinx.net.za - 0 | 129 | 129 | 159 | 162 | 231 | 160 |
| hosted-by.i3d.net - 0 | 129 | 129 | 159 | 160 | 206 | 161 |
|________________________________________________|______|______|______|______|______|______|
WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider


Friend's WinMTR to same local server via competing local ISP:

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| dlinkrouter - 0 | 1048 | 1048 | 0 | 0 | 15 | 0 |
| 102-182-101-1.ip.afrihost.joburg - 0 | 1048 | 1048 | 2 | 5 | 123 | 2 |
| lucopel.net.afrihost.co.za - 0 | 1049 | 1049 | 1 | 1 | 12 | 1 |
| 169-1-21-91.ip.afrihost.co.za - 0 | 1048 | 1048 | 1 | 1 | 29 | 1 |
| i3d.ixp.joburg - 0 | 1048 | 1048 | 1 | 3 | 66 | 2 |
| hosted-by.i3d.net - 0 | 1048 | 1048 | 1 | 1 | 16 | 1 |
|________________________________________________|______|______|______|______|______|______|

WinMTR v1.00 GPLv2 (original by Appnor MSP - Fully Managed Hosting & Cloud Provider)

Edited post: IP used to test against, Rocket League server @ 185.179.201.74
 
Last edited:

TheRoDent

Cool Ideas Rep
Joined
Aug 6, 2003
Messages
5,695
They allowed you to play on any region so for some odd region the Asian hackers at the time came onto the EU server to ruin their games, the only way they could control it was to apply a region lock.
How would a region lock control game hacking, if they are in control of the server in any case? This just smells of a badly implemented game server, and/or client.

Server regionality will never be able to help in the scenario of "hacking". In fact community servers will probably do better, since communities tend to self-moderate.

A server should never trust a client, but it's also nigh impossible to control what a client does with memory hacking or injection. But a good game should never trust a client, and allow them to do extraordinary things, like warping instantly, or speedhacking. Client-side "reveal" hacking will always be a problem, but in a good game design it shouldn't matter when it comes down to melee. Anyway, this could be a very long discussion.... I've had these kind of conversations about FPS "trust" for more than 15 years.

And yes, I've built games. FPS ones. And have had to deal with all of this.


1622763016373.png
 

Attachments

  • 1622762902916.png
    1622762902916.png
    77.6 KB · Views: 17
Last edited:

Jason-ZA

Well-Known Member
Joined
Sep 23, 2019
Messages
333
Blizzard seems to have fixed their routing, ping back to normal again

Code:
|------------------------------------------------------------------------------------------|
|                                      WinMTR statistics                                   |
|                       Host              -   %  | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
|                                10.0.0.1 -    0 |    5 |    5 |    0 |    0 |    0 |    0 |
|                             154.0.0.246 -    0 |    5 |    5 |   11 |   11 |   12 |   11 |
|                              100.98.0.3 -    0 |    5 |    5 |   11 |   11 |   12 |   11 |
|                            100.99.0.126 -    0 |    5 |    5 |  150 |  150 |  151 |  151 |
|                             100.98.1.98 -    0 |    5 |    5 |  151 |  151 |  151 |  151 |
|                          195.66.226.234 -    0 |    5 |    5 |  152 |  152 |  153 |  153 |
|                           137.221.79.33 -    0 |    5 |    5 |  158 |  176 |  226 |  158 |
|                   No response from host -    0 |    0 |    0 |    0 |    0 |    0 |    0 |
|                           137.221.65.77 -    0 |    5 |    5 |  158 |  194 |  340 |  158 |
|                           137.221.78.69 -    0 |    5 |    5 |  158 |  158 |  158 |  158 |
|                           137.221.66.47 -    0 |    5 |    5 |  158 |  158 |  158 |  158 |
|                          185.60.112.157 -    0 |    5 |    5 |  158 |  158 |  158 |  158 |
|________________________________________________|______|______|______|______|______|______|
   WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider
 

TheRoDent

Cool Ideas Rep
Joined
Aug 6, 2003
Messages
5,695
How would a region lock control game hacking, if they are in control of the server in any case? This just smells of a badly implemented game server.

Server regionality will never be able to help in the scenario of "hacking". In fact community servers will probably do better, since communities tend to self-moderate.

A server should never trust a client, but it's also nigh impossible to control what a client does with memory hacking or injection. But a good game should never trust a client, and allow them to do extraordinary things, like warping instantly, or speedhacking. Client-side "reveal" hacking will always be a problem, but in a good game design it shouldn't matter when it comes down to melee. Anyway, this could be a very long discussion.... I've had these kind of conversations about FPS "trust" for more than 15 years.

And yes, I've built games. FPS ones. And have had to deal with all of this.


View attachment 1082839


Also, http://rodent.za.net/defenseturret/index.html

1622763375077.png

This was many moons ago, but Tribes 1/2, was one of the best community supported, and moddable games, and whilst hacking was an issue, in the end, in competive play, cheating wouldn't make a major difference.

Tribes 2 invented e-sports, and "casting/streaming". Justin.tv (which became Twitch) was founded by Emmett because of his passion for e-sports, and Tribes, and was part of the modding community that drove Tribes2, casting, and e-sports. Some of the very first justin.tv events were casted from T2 games.

Read this article about "Before Fortnite".

This was in the days when I still messed with assembly, game dev, and everything around it.

Now I just push packets and cry when I hear about where game servers are going, and how game servers are monopolised.
 
Last edited:

TheRoDent

Cool Ideas Rep
Joined
Aug 6, 2003
Messages
5,695
@PBCool I mean no disrespect but it does not make sense, please explain? It is the fourth day that I get "international" 170ms instead of 20ms ping.

All my friends on competing local ISPs has no such issue. And on our Rocket League group I got no confirmation from anyone else having this issue. Please have a look at the following two WinMTRs.

Here is tonight's WinMTR via my Cool Ideas account (versus competing ISP after):

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| 192.168.1.1 - 0 | 129 | 129 | 0 | 0 | 1 | 0 |
| u6q-cust.coolideas.co.za - 0 | 129 | 129 | 1 | 2 | 37 | 2 |
| 100.98.0.1 - 0 | 129 | 129 | 1 | 2 | 38 | 2 |
| 100.98.0.49 - 0 | 129 | 129 | 16 | 17 | 52 | 18 |
| 100.99.0.180 - 0 | 129 | 129 | 16 | 17 | 51 | 18 |
| i3d.jinx.net.za - 0 | 129 | 129 | 159 | 162 | 231 | 160 |
| hosted-by.i3d.net - 0 | 129 | 129 | 159 | 160 | 206 | 161 |
|________________________________________________|______|______|______|______|______|______|
WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider


Friend's WinMTR to same local server via competing local ISP:

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| dlinkrouter - 0 | 1048 | 1048 | 0 | 0 | 15 | 0 |
| 102-182-101-1.ip.afrihost.joburg - 0 | 1048 | 1048 | 2 | 5 | 123 | 2 |
| lucopel.net.afrihost.co.za - 0 | 1049 | 1049 | 1 | 1 | 12 | 1 |
| 169-1-21-91.ip.afrihost.co.za - 0 | 1048 | 1048 | 1 | 1 | 29 | 1 |
| i3d.ixp.joburg - 0 | 1048 | 1048 | 1 | 3 | 66 | 2 |
| hosted-by.i3d.net - 0 | 1048 | 1048 | 1 | 1 | 16 | 1 |
|________________________________________________|______|______|______|______|______|______|

WinMTR v1.00 GPLv2 (original by Appnor MSP - Fully Managed Hosting & Cloud Provider)

Edited post: IP used to test against, Rocket League server @ 185.179.201.74
So, the issue is that i3d has pulled that server back to a location somewhere far away, or they have some strange return routing, but they are still advertising the IP at local exchanges, such as JINX.

We hand over traffic to them from our network at JINX, and from there we are not entirely in control from that point further on.

They made a change, and we didn't, so it's worse now, but it's not our fault, if that makes sense ?
I would appreciate if you can PM me your IP address, so I can just check the source as well.

However it seems from the Afrihost trace that the NAPAfrica path is better, which probably means i3d is neglecting their JINX peering, so I will chat to our BGP engineers to rather prefer the NAP route, which brings me to my next point...

I'm really starting to dislike JINX (not specifically for this reason), and think that as a future policy we're just going to generally weight JINX lower. It just seems like in general peers care more about their NAP activities, than JINX.

We will fix this for you.
 
Last edited:

TheRoDent

Cool Ideas Rep
Joined
Aug 6, 2003
Messages
5,695
@PBCool I mean no disrespect but it does not make sense, please explain? It is the fourth day that I get "international" 170ms instead of 20ms ping.

All my friends on competing local ISPs has no such issue. And on our Rocket League group I got no confirmation from anyone else having this issue. Please have a look at the following two WinMTRs.

Here is tonight's WinMTR via my Cool Ideas account (versus competing ISP after):

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| 192.168.1.1 - 0 | 129 | 129 | 0 | 0 | 1 | 0 |
| u6q-cust.coolideas.co.za - 0 | 129 | 129 | 1 | 2 | 37 | 2 |
| 100.98.0.1 - 0 | 129 | 129 | 1 | 2 | 38 | 2 |
| 100.98.0.49 - 0 | 129 | 129 | 16 | 17 | 52 | 18 |
| 100.99.0.180 - 0 | 129 | 129 | 16 | 17 | 51 | 18 |
| i3d.jinx.net.za - 0 | 129 | 129 | 159 | 162 | 231 | 160 |
| hosted-by.i3d.net - 0 | 129 | 129 | 159 | 160 | 206 | 161 |
|________________________________________________|______|______|______|______|______|______|
WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider


Friend's WinMTR to same local server via competing local ISP:

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| dlinkrouter - 0 | 1048 | 1048 | 0 | 0 | 15 | 0 |
| 102-182-101-1.ip.afrihost.joburg - 0 | 1048 | 1048 | 2 | 5 | 123 | 2 |
| lucopel.net.afrihost.co.za - 0 | 1049 | 1049 | 1 | 1 | 12 | 1 |
| 169-1-21-91.ip.afrihost.co.za - 0 | 1048 | 1048 | 1 | 1 | 29 | 1 |
| i3d.ixp.joburg - 0 | 1048 | 1048 | 1 | 3 | 66 | 2 |
| hosted-by.i3d.net - 0 | 1048 | 1048 | 1 | 1 | 16 | 1 |
|________________________________________________|______|______|______|______|______|______|

WinMTR v1.00 GPLv2 (original by Appnor MSP - Fully Managed Hosting & Cloud Provider)

Edited post: IP used to test against, Rocket League server @ 185.179.201.74
This seems to have returned to local pings, but we will still investigate better routing, and prefer NAP.
 

leppie

Senior Member
Joined
Mar 25, 2019
Messages
527
Yeah, the -O2 omits some of the first part of the test, which is kinda normal for allowing rampup etc.

@leppie can you maybe run a longer test (-t 60) and -O2 (omit first two)?

That was all fine, but I think I see the issue. The out of order is the hint. I am getting duplicate packets at times.

Why would I be getting the same packet 3 times? I assume in TCP, this will cause a retry.

Code:
iperf3: OUT OF ORDER - incoming packet = 1793 and received packet = 1801 AND SP = 4
iperf3: OUT OF ORDER - incoming packet = 1794 and received packet = 1802 AND SP = 4
iperf3: OUT OF ORDER - incoming packet = 1795 and received packet = 1802 AND SP = 4
iperf3: OUT OF ORDER - incoming packet = 1796 and received packet = 1802 AND SP = 4
[  4]   3.00-4.00   sec  2.38 MBytes  19.9 Mbits/sec  0.607 ms  4/304 (1.3%)
 

CrypticZA

Expert Member
Joined
Sep 21, 2019
Messages
1,934
How would a region lock control game hacking, if they are in control of the server in any case? This just smells of a badly implemented game server, and/or client.

Server regionality will never be able to help in the scenario of "hacking". In fact community servers will probably do better, since communities tend to self-moderate.

A server should never trust a client, but it's also nigh impossible to control what a client does with memory hacking or injection. But a good game should never trust a client, and allow them to do extraordinary things, like warping instantly, or speedhacking. Client-side "reveal" hacking will always be a problem, but in a good game design it shouldn't matter when it comes down to melee. Anyway, this could be a very long discussion.... I've had these kind of conversations about FPS "trust" for more than 15 years.

And yes, I've built games. FPS ones. And have had to deal with all of this.


View attachment 1082839
Its not to control hacking but more to keep it in a region and to stop rage hacking so in the case of PUBG the Asian countries were apparently hacking on purpose in the EU lobbies because of something that happened I can't remember what but in the end it was like a "we don't like you we are going to make sure you don't have fun in this game." I agree with the you should play wherever you want but some game develops also say its not fair if you have someone on your team playing a competitive game with an extreme high ping it ruins it for the rest of the team, however you could work around this and have casual games selecting any region and competitive force a certain server.

So the games I play mainly Dota 2 and Apex allow you to select any region you want, however Apex within the last few months has grown so much that instead of choosing a specific server to play on you get assigned a region. So Belgium/Amsterdam/London/Frankfurt are all EU so you will get any of those when you select one of them and the idea is to get people into matches quicker. In their world everyone from one region should have ideally more or less the same ping to all the servers in a certain region, however with the likes of us, Google and Amazon transit screw this ideal world up for us.

Using a UK VPN the pings range from 145 to 155 however off VPN due to Google and AWS they range from 145 to 200.
 

MDKza

Senior Member
Joined
May 24, 2012
Messages
726
What do you mean "second" connection ?

I have a session up for you on our side. If you have an alternative "second" connection on your router that is not with us, perhaps that is taking priority. I don't understand your setup, but welcome to continue the conversation in our PM.

and suddenly it's up again - without changing a thing.
 

DYreX146

Senior Member
Joined
Nov 14, 2018
Messages
511
Which is wrong. You should be able to choose with whom, and where you want to play.

Gaming is a social thing. Not a ping/slot-decided-by-a-gaming-company thing.
Same thing happened with Escape From Tarkov.

Many people in EU and NA complained that people on their servers from other regions were "ping abusing" and causing them to lose fights because our latency was much higher.

So the devs decided to try keep everyone happy by giving us SA servers and creating a ping lock. Yet even now those same people complain when they lose because of desync with the server, so if it's not ping abuse now it's desync.

Ping abuse was just an excuse in my opinion because some very skilled players in the SA community, one of whom has over 10 000 hours in the game, agree that such a thing doesn't exist and it's just desync that everyone suffers from regardless of latency.

Because of this I now can't play with my German friends who helped me learn the game when I started playing 3 years ago since you get kicked from the server if you have over 165ms. Our servers are basically a PvE simulator since there's probably only 20 people playing at any given time.
 

CrypticZA

Expert Member
Joined
Sep 21, 2019
Messages
1,934
Same thing happened with Escape From Tarkov.

Many people in EU and NA complained that people on their servers from other regions were "ping abusing" and causing them to lose fights because our latency was much higher.

So the devs decided to try keep everyone happy by giving us SA servers and creating a ping lock. Yet even now those same people complain when they lose because of desync with the server, so if it's not ping abuse now it's desync.

Ping abuse was just an excuse in my opinion because some very skilled players in the SA community, one of whom has over 10 000 hours in the game, agree that such a thing doesn't exist and it's just desync that everyone suffers from regardless of latency.

Because of this I now can't play with my German friends who helped me learn the game when I started playing 3 years ago since you get kicked from the server if you have over 165ms. Our servers are basically a PvE simulator since there's probably only 20 people playing at any given time.
Use to get kicked all the time from Battlefield servers because of ping...

If you from CT you should be able to get 155 to Germany but i think JHB is 165 is the lowest
 

DYreX146

Senior Member
Joined
Nov 14, 2018
Messages
511
Use to get kicked all the time from Battlefield servers because of ping...

If you from CT you should be able to get 155 to Germany but i think JHB is 165 is the lowest
I'm in KZN so it's about 180ms at least. Only CPT people can play EFT on EU but even they struggle often because the servers fluctuate up to 20ms extra very often.

This is what my server list looks like:

1622793244655.png
 

DeatheCore

Senior Member
Joined
Dec 4, 2006
Messages
795
That was all fine, but I think I see the issue. The out of order is the hint. I am getting duplicate packets at times.

Why would I be getting the same packet 3 times? I assume in TCP, this will cause a retry.

Code:
iperf3: OUT OF ORDER - incoming packet = 1793 and received packet = 1801 AND SP = 4
iperf3: OUT OF ORDER - incoming packet = 1794 and received packet = 1802 AND SP = 4
iperf3: OUT OF ORDER - incoming packet = 1795 and received packet = 1802 AND SP = 4
iperf3: OUT OF ORDER - incoming packet = 1796 and received packet = 1802 AND SP = 4
[  4]   3.00-4.00   sec  2.38 MBytes  19.9 Mbits/sec  0.607 ms  4/304 (1.3%)
You're using iperf 3.1.3 right? Would suggest using a newer version and testing again. There have been significant bugfixes and changes since then, including changes to UDP tests and reporting (no longer defaults to 8K blocks which lead to fragmentation).

Looks like BudMan has compiled v3.10.1 for Windows - https://files.budman.pw/iperf3.10.1_64bit.zip
 
Last edited:

leppie

Senior Member
Joined
Mar 25, 2019
Messages
527
You're using iperf 3.1.3 right? Would suggest using a newer version and testing again. There have been significant bugfixes and changes since then, including changes to UDP tests and reporting (no longer defaults to 8K blocks which lead to fragmentation).

Looks like BudMan has compiled 3.10 for Windows - https://files.budman.pw/iperf3.10_64bit.zip

Thanks.

@TheRoDent Here is 120 second trace. Every 5-15 seconds, there is a notable 3% loss. I still believe this is caused by the out of order packets that is sending back duplicates. 0 loss lines omitted.

Code:
iperf3 -R -u -b 20M -O 2 -t 120 -p 17001 -c cptspeedtest.cisp.co.za
Connecting to host cptspeedtest.cisp.co.za, port 17001
Reverse mode, remote host cptspeedtest.cisp.co.za is sending
[  5] local 192.168.0.43 port 54433 connected to 154.0.15.181 port 17001
[ ID] Interval           Transfer     Bitrate         Jitter    Lost/Total Datagrams
[  5]   0.00-1.00   sec  2.45 MBytes  20.5 Mbits/sec  0.037 ms  35/1818 (1.9%)  (omitted)
[  5]   1.00-2.00   sec  2.38 MBytes  20.0 Mbits/sec  0.157 ms  0/1736 (0%)  (omitted)
[  5]   1.00-2.00   sec  2.36 MBytes  19.8 Mbits/sec  0.057 ms  18/1735 (1%)
[  5]   7.00-8.00   sec  2.27 MBytes  19.0 Mbits/sec  0.087 ms  88/1738 (5.1%)
[  5]  15.00-16.00  sec  2.35 MBytes  19.7 Mbits/sec  0.040 ms  23/1736 (1.3%)
[  5]  16.00-17.00  sec  2.38 MBytes  20.0 Mbits/sec  0.066 ms  1/1736 (0.058%)
[  5]  21.00-22.00  sec  2.29 MBytes  19.2 Mbits/sec  0.072 ms  67/1735 (3.9%)
[  5]  31.00-32.00  sec  2.32 MBytes  19.5 Mbits/sec  0.102 ms  44/1736 (2.5%)
[  5]  40.00-41.00  sec  2.36 MBytes  19.8 Mbits/sec  0.143 ms  21/1736 (1.2%)
[  5]  41.00-42.00  sec  2.38 MBytes  19.9 Mbits/sec  0.145 ms  5/1736 (0.29%)
[  5]  46.00-47.00  sec  2.28 MBytes  19.1 Mbits/sec  0.207 ms  77/1736 (4.4%)
[  5]  55.00-56.00  sec  2.37 MBytes  19.9 Mbits/sec  0.073 ms  10/1736 (0.58%)
[  5]  77.00-78.00  sec  2.31 MBytes  19.4 Mbits/sec  0.091 ms  54/1736 (3.1%)
[  5]  86.00-87.00  sec  2.31 MBytes  19.4 Mbits/sec  0.082 ms  51/1736 (2.9%)
[  5]  95.00-96.00  sec  2.31 MBytes  19.4 Mbits/sec  0.161 ms  55/1734 (3.2%)
[  5] 106.00-107.00 sec  2.25 MBytes  18.9 Mbits/sec  0.209 ms  97/1737 (5.6%)
[  5] 107.00-108.00 sec  2.38 MBytes  20.0 Mbits/sec  6.262 ms  5/1735 (0.29%)
[  5] 108.00-109.00 sec  2.39 MBytes  20.1 Mbits/sec  0.091 ms  -5/1736 (-0.29%)
[  5] 115.00-116.00 sec  2.31 MBytes  19.4 Mbits/sec  0.324 ms  48/1730 (2.8%)
[  5] 116.00-117.00 sec  2.37 MBytes  19.9 Mbits/sec  0.079 ms  18/1742 (1%)
[  5] 119.00-120.00 sec  2.39 MBytes  20.0 Mbits/sec  0.095 ms  0/1737 (0%)
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Jitter    Lost/Total Datagrams
[  5]   0.00-120.05 sec   286 MBytes  20.0 Mbits/sec  0.000 ms  0/208341 (0%)  sender
[SUM]  0.0-120.1 sec  472 datagrams received out-of-order
[  5]   0.00-120.00 sec   285 MBytes  19.9 Mbits/sec  0.095 ms  677/208337 (0.32%)  receiver
 

jannier

Expert Member
Joined
Jul 31, 2005
Messages
1,836
Thanks.

@TheRoDent Here is 120 second trace. Every 5-15 seconds, there is a notable 3% loss. I still believe this is caused by the out of order packets that is sending back duplicates. 0 loss lines omitted.

Code:
iperf3 -R -u -b 20M -O 2 -t 120 -p 17001 -c cptspeedtest.cisp.co.za
Connecting to host cptspeedtest.cisp.co.za, port 17001
Reverse mode, remote host cptspeedtest.cisp.co.za is sending
[  5] local 192.168.0.43 port 54433 connected to 154.0.15.181 port 17001
[ ID] Interval           Transfer     Bitrate         Jitter    Lost/Total Datagrams
[  5]   0.00-1.00   sec  2.45 MBytes  20.5 Mbits/sec  0.037 ms  35/1818 (1.9%)  (omitted)
[  5]   1.00-2.00   sec  2.38 MBytes  20.0 Mbits/sec  0.157 ms  0/1736 (0%)  (omitted)
[  5]   1.00-2.00   sec  2.36 MBytes  19.8 Mbits/sec  0.057 ms  18/1735 (1%)
[  5]   7.00-8.00   sec  2.27 MBytes  19.0 Mbits/sec  0.087 ms  88/1738 (5.1%)
[  5]  15.00-16.00  sec  2.35 MBytes  19.7 Mbits/sec  0.040 ms  23/1736 (1.3%)
[  5]  16.00-17.00  sec  2.38 MBytes  20.0 Mbits/sec  0.066 ms  1/1736 (0.058%)
[  5]  21.00-22.00  sec  2.29 MBytes  19.2 Mbits/sec  0.072 ms  67/1735 (3.9%)
[  5]  31.00-32.00  sec  2.32 MBytes  19.5 Mbits/sec  0.102 ms  44/1736 (2.5%)
[  5]  40.00-41.00  sec  2.36 MBytes  19.8 Mbits/sec  0.143 ms  21/1736 (1.2%)
[  5]  41.00-42.00  sec  2.38 MBytes  19.9 Mbits/sec  0.145 ms  5/1736 (0.29%)
[  5]  46.00-47.00  sec  2.28 MBytes  19.1 Mbits/sec  0.207 ms  77/1736 (4.4%)
[  5]  55.00-56.00  sec  2.37 MBytes  19.9 Mbits/sec  0.073 ms  10/1736 (0.58%)
[  5]  77.00-78.00  sec  2.31 MBytes  19.4 Mbits/sec  0.091 ms  54/1736 (3.1%)
[  5]  86.00-87.00  sec  2.31 MBytes  19.4 Mbits/sec  0.082 ms  51/1736 (2.9%)
[  5]  95.00-96.00  sec  2.31 MBytes  19.4 Mbits/sec  0.161 ms  55/1734 (3.2%)
[  5] 106.00-107.00 sec  2.25 MBytes  18.9 Mbits/sec  0.209 ms  97/1737 (5.6%)
[  5] 107.00-108.00 sec  2.38 MBytes  20.0 Mbits/sec  6.262 ms  5/1735 (0.29%)
[  5] 108.00-109.00 sec  2.39 MBytes  20.1 Mbits/sec  0.091 ms  -5/1736 (-0.29%)
[  5] 115.00-116.00 sec  2.31 MBytes  19.4 Mbits/sec  0.324 ms  48/1730 (2.8%)
[  5] 116.00-117.00 sec  2.37 MBytes  19.9 Mbits/sec  0.079 ms  18/1742 (1%)
[  5] 119.00-120.00 sec  2.39 MBytes  20.0 Mbits/sec  0.095 ms  0/1737 (0%)
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Jitter    Lost/Total Datagrams
[  5]   0.00-120.05 sec   286 MBytes  20.0 Mbits/sec  0.000 ms  0/208341 (0%)  sender
[SUM]  0.0-120.1 sec  472 datagrams received out-of-order
[  5]   0.00-120.00 sec   285 MBytes  19.9 Mbits/sec  0.095 ms  677/208337 (0.32%)  receiver

You have loss on your connection. I ran the same for comparison.

Code:
c:\iperf3>iperf3 -R -u -b 20M -O2 -t 120 -p 17001 -c cptspeedtest.cisp.co.za
Connecting to host cptspeedtest.cisp.co.za, port 17001
Reverse mode, remote host cptspeedtest.cisp.co.za is sending
[  5] local 192.168.88.213 port 53757 connected to 154.0.15.181 port 17001
[ ID] Interval           Transfer     Bitrate         Jitter    Lost/Total Datagrams
[  5]   0.00-1.00   sec  2.45 MBytes  20.5 Mbits/sec  0.094 ms  22/1780 (1.2%)  (omitted)
[  5]   1.00-2.00   sec  2.38 MBytes  20.0 Mbits/sec  0.077 ms  0/1712 (0%)  (omitted)
[  5]   0.00-1.00   sec  2.39 MBytes  20.0 Mbits/sec  0.052 ms  0/1714 (0%)
[  5]   1.00-2.00   sec  2.38 MBytes  20.0 Mbits/sec  0.068 ms  0/1712 (0%)
[  5]   2.00-3.00   sec  2.39 MBytes  20.0 Mbits/sec  0.061 ms  0/1714 (0%)
[  5]   3.00-4.00   sec  2.38 MBytes  20.0 Mbits/sec  0.105 ms  0/1711 (0%)
[  5]   4.00-5.00   sec  2.38 MBytes  20.0 Mbits/sec  0.114 ms  0/1712 (0%)
[  5]   5.00-6.00   sec  2.39 MBytes  20.0 Mbits/sec  0.057 ms  0/1714 (0%)
[  5]   6.00-7.00   sec  2.38 MBytes  20.0 Mbits/sec  0.069 ms  0/1712 (0%)
[  5]   7.00-8.00   sec  2.38 MBytes  20.0 Mbits/sec  0.074 ms  0/1711 (0%)
[  5]   8.00-9.00   sec  2.39 MBytes  20.0 Mbits/sec  0.048 ms  0/1714 (0%)
[  5]   9.00-10.00  sec  2.38 MBytes  20.0 Mbits/sec  0.082 ms  0/1710 (0%)
[  5]  10.00-11.00  sec  2.39 MBytes  20.0 Mbits/sec  0.069 ms  0/1713 (0%)
[  5]  11.00-12.00  sec  2.39 MBytes  20.0 Mbits/sec  0.058 ms  0/1713 (0%)
[  5]  12.00-13.00  sec  2.39 MBytes  20.0 Mbits/sec  0.055 ms  0/1713 (0%)
[  5]  13.00-14.00  sec  2.38 MBytes  20.0 Mbits/sec  0.095 ms  0/1711 (0%)
[  5]  14.00-15.00  sec  2.39 MBytes  20.0 Mbits/sec  0.083 ms  0/1713 (0%)
[  5]  15.00-16.00  sec  2.38 MBytes  20.0 Mbits/sec  0.060 ms  0/1712 (0%)
removed some lines to fit the 1000 character limit on post.
[  5] 115.00-116.00 sec  2.38 MBytes  20.0 Mbits/sec  0.062 ms  0/1712 (0%)
[  5] 116.00-117.00 sec  2.39 MBytes  20.0 Mbits/sec  0.080 ms  0/1714 (0%)
[  5] 117.00-118.00 sec  2.38 MBytes  20.0 Mbits/sec  0.095 ms  0/1712 (0%)
[  5] 118.00-119.00 sec  2.38 MBytes  20.0 Mbits/sec  0.048 ms  0/1711 (0%)
[  5] 119.00-120.00 sec  2.38 MBytes  20.0 Mbits/sec  0.065 ms  0/1712 (0%)
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bitrate         Jitter    Lost/Total Datagrams
[  5]   0.00-120.04 sec   286 MBytes  20.0 Mbits/sec  0.000 ms  0/205483 (0%)  sender
[  5]   0.00-120.00 sec   286 MBytes  20.0 Mbits/sec  0.065 ms  0/205479 (0%)  receiver

iperf Done.
 
Top