Cool Ideas Fibre ISP - Feedback

Status
Not open for further replies.
Sinbad said:
|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| Keybaud - 10 | 968 | 876 | 0 | 0 | 9 | 1 |
| 155.93.152.1 - 0 | 1797 | 1797 | 0 | 1 | 172 | 1 |
| 154.0.3.97 - 1 | 1745 | 1739 | 2 | 7 | 168 | 12 |
| 154.0.2.1 - 1 | 1718 | 1709 | 0 | 0 | 36 | 1 |
| 154.0.4.54 - 1 | 1754 | 1749 | 1 | 1 | 30 | 1 |
| 154.0.4.254 - 1 | 1694 | 1682 | 158 | 160 | 368 | 163 |
| 195.66.226.234 - 1 | 1702 | 1691 | 159 | 164 | 397 | 161 |
| ae1-br01-eqld5.blizzardonline.net - 1 | 1671 | 1658 | 160 | 176 | 1520 | 164 |
| et-0-0-2-br02-eqam1.blizzardonline.net - 1 | 1710 | 1700 | 166 | 175 | 428 | 168 |
| et-0-0-67-pe01-eqam1.blizzardonline.net - 1 | 1709 | 1699 | 165 | 171 | 558 | 169 |
| 137.221.66.41 - 1 | 1684 | 1671 | 165 | 171 | 473 | 168 |
| 10.105.1.34 - 1 | 1728 | 1720 | 165 | 171 | 382 | 168 |
| 37.244.54.10 - 1 | 1693 | 1681 | 165 | 172 | 532 | 168 |
|________________________________________________|______|______|______|______|______|______|
WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider
Click to expand...
Thanks I will run the same from my place to compare.
 
My connection to my AWS instance in London has been flapping as of late.

Just can't seem to keep a throughput higher than 4mbps (not MB/s) to my server. It's starting to become rather annoying that i can't keep a stable connection of 4% of my line capacity.

Doesn't seem like a routing issue either.

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| 172.18.7.225 - 0 | 2115 | 2115 | 0 | 0 | 3 | 0 |
| 155.93.246.1 - 1 | 2104 | 2101 | 0 | 1 | 118 | 0 |
| 154.0.1.125 - 0 | 2115 | 2115 | 0 | 0 | 31 | 0 |
| 154.0.1.13 - 0 | 2115 | 2115 | 1 | 1 | 62 | 1 |
| 154.0.3.106 - 0 | 2115 | 2115 | 141 | 141 | 367 | 141 |
| 195.66.225.175 - 1 | 2111 | 2110 | 141 | 143 | 388 | 142 |
| No response from host - 100 | 424 | 0 | 0 | 0 | 0 | 0 |
| 52.94.35.57 - 1 | 2111 | 2110 | 141 | 142 | 192 | 142 |
| 52.94.35.16 - 0 | 2115 | 2115 | 143 | 155 | 398 | 160 |
| 52.94.33.135 - 0 | 2115 | 2115 | 142 | 143 | 442 | 144 |
| 52.94.33.38 - 0 | 2115 | 2115 | 142 | 143 | 347 | 143 |
| No response from host - 100 | 424 | 0 | 0 | 0 | 0 | 0 |
| No response from host - 100 | 424 | 0 | 0 | 0 | 0 | 0 |
| No response from host - 100 | 424 | 0 | 0 | 0 | 0 | 0 |
|ec2-35-176-155-105.eu-west-2.compute.amazonaws.com - 1 | 2111 | 2110 | 142 | 143 | 488 | 143 |
|________________________________________________|______|______|______|______|______|______|
WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider
 
DJZassie said:
My connection to my AWS instance in London has been flapping as of late.

Just can't seem to keep a throughput higher than 4mbps (not MB/s) to my server. It's starting to become rather annoying that i can't keep a stable connection of 4% of my line capacity.

Doesn't seem like a routing issue either.

|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| 172.18.7.225 - 0 | 2115 | 2115 | 0 | 0 | 3 | 0 |
| 155.93.246.1 - 1 | 2104 | 2101 | 0 | 1 | 118 | 0 |
| 154.0.1.125 - 0 | 2115 | 2115 | 0 | 0 | 31 | 0 |
| 154.0.1.13 - 0 | 2115 | 2115 | 1 | 1 | 62 | 1 |
| 154.0.3.106 - 0 | 2115 | 2115 | 141 | 141 | 367 | 141 |
| 195.66.225.175 - 1 | 2111 | 2110 | 141 | 143 | 388 | 142 |
| No response from host - 100 | 424 | 0 | 0 | 0 | 0 | 0 |
| 52.94.35.57 - 1 | 2111 | 2110 | 141 | 142 | 192 | 142 |
| 52.94.35.16 - 0 | 2115 | 2115 | 143 | 155 | 398 | 160 |
| 52.94.33.135 - 0 | 2115 | 2115 | 142 | 143 | 442 | 144 |
| 52.94.33.38 - 0 | 2115 | 2115 | 142 | 143 | 347 | 143 |
| No response from host - 100 | 424 | 0 | 0 | 0 | 0 | 0 |
| No response from host - 100 | 424 | 0 | 0 | 0 | 0 | 0 |
| No response from host - 100 | 424 | 0 | 0 | 0 | 0 | 0 |
|ec2-35-176-155-105.eu-west-2.compute.amazonaws.com - 1 | 2111 | 2110 | 142 | 143 | 488 | 143 |
|________________________________________________|______|______|______|______|______|______|
WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider
Click to expand...

Yeah looks fine, have you logged it with AWS?
 
Sinbad said:
|------------------------------------------------------------------------------------------|
| WinMTR statistics |
| Host - % | Sent | Recv | Best | Avrg | Wrst | Last |
|------------------------------------------------|------|------|------|------|------|------|
| Keybaud - 10 | 968 | 876 | 0 | 0 | 9 | 1 |
| 155.93.152.1 - 0 | 1797 | 1797 | 0 | 1 | 172 | 1 |
| 154.0.3.97 - 1 | 1745 | 1739 | 2 | 7 | 168 | 12 |
| 154.0.2.1 - 1 | 1718 | 1709 | 0 | 0 | 36 | 1 |
| 154.0.4.54 - 1 | 1754 | 1749 | 1 | 1 | 30 | 1 |
| 154.0.4.254 - 1 | 1694 | 1682 | 158 | 160 | 368 | 163 |
| 195.66.226.234 - 1 | 1702 | 1691 | 159 | 164 | 397 | 161 |
| ae1-br01-eqld5.blizzardonline.net - 1 | 1671 | 1658 | 160 | 176 | 1520 | 164 |
| et-0-0-2-br02-eqam1.blizzardonline.net - 1 | 1710 | 1700 | 166 | 175 | 428 | 168 |
| et-0-0-67-pe01-eqam1.blizzardonline.net - 1 | 1709 | 1699 | 165 | 171 | 558 | 169 |
| 137.221.66.41 - 1 | 1684 | 1671 | 165 | 171 | 473 | 168 |
| 10.105.1.34 - 1 | 1728 | 1720 | 165 | 171 | 382 | 168 |
| 37.244.54.10 - 1 | 1693 | 1681 | 165 | 172 | 532 | 168 |
|________________________________________________|______|______|______|______|______|______|
WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider
Click to expand...

I know it might not help much, but here is mine on Vodacom 100/50mbps.
winmtr.jpg
 
PBCool said:
Yeah looks fine, have you logged it with AWS?
Click to expand...

Indeed,

We tried stopping my server and starting it again (This would force the server onto new hardware as it is a cloud server)
We tried spinning it up on higher rated hardware (up to 10 gigabit network ports to make sure that wasn't the issue)

And alas, the same issue both times.

It a case of the connection holds perfectly fine for anywhere from 5 to 30 minutes, then gets very very dodgy for 1 to 5 minutes and then goes fine again. By dodgy i mean it pushes data at a much slower rate for that period of time and then picks up again.
 
Dear Cool Ideas,
I'm no longer able to tolerate the heavy packet loss on your international leg of your route due to your current contention construct.
I understand that you capacity manage at 80% but this depends on your monitoring intervals and therefore you have a degree of acceptable packet drops.
Unfortunately the applications that I use and your degree of packet loss tolerance is incompatible and even though I have requested that my service (20/20) be changed to another service category (many many times) in order to ascertain if this will improve my experience. I was hoping that you either have a different international allocation for your premium services or CoS hierarchy.
I understand that the market for consumer broadband is driven by media consumption and therefore caching would suffice is most cases but there are more serious users (home office) like me that require better access to non-traditional traffic sources (large public cloud providers) internationally and therefore the generic support process will continue to yield this type of customer churn.
 
azzkikr1 said:
Dear Cool Ideas,
I'm no longer able to tolerate the heavy packet loss on your international leg of your route due to your current contention construct.
I understand that you capacity manage at 80% but this depends on your monitoring intervals and therefore you have a degree of acceptable packet drops.
Unfortunately the applications that I use and your degree of packet loss tolerance is incompatible and even though I have requested that my service (20/20) be changed to another service category (many many times) in order to ascertain if this will improve my experience. I was hoping that you either have a different international allocation for your premium services or CoS hierarchy.
I understand that the market for consumer broadband is driven by media consumption and therefore caching would suffice is most cases but there are more serious users (home office) like me that require better access to non-traditional traffic sources (large public cloud providers) internationally and therefore the generic support process will continue to yield this type of customer churn.
Click to expand...

Hi do you have a ticket number for me to investigate, we manage capacities at 80% on a live basis, not on a polling interval.
 
DJZassie said:
Indeed,

We tried stopping my server and starting it again (This would force the server onto new hardware as it is a cloud server)
We tried spinning it up on higher rated hardware (up to 10 gigabit network ports to make sure that wasn't the issue)

And alas, the same issue both times.

It a case of the connection holds perfectly fine for anywhere from 5 to 30 minutes, then gets very very dodgy for 1 to 5 minutes and then goes fine again. By dodgy i mean it pushes data at a much slower rate for that period of time and then picks up again.
Click to expand...

If i'm not mistaken you had the same results when we were routing via the AWS Direct connect?
 
PBCool said:
If i'm not mistaken you had the same results when we were routing via the AWS Direct connect?
Click to expand...

You are correct. The direct connect was terrible so you set me up to go via your peer in London.
That worked great until now where it seems to be very similar to how it was for the direct connect.

But i'm 90% sure this isn't an issue with AWS as i have had mates on different ISP's try and use my AWS Instance and they have not had the same issue as me.
 
Bedford said:
twitch has always been an issue and its a known issue. on my 50/50 line I very rarely can watch at 1080p sometimes even 720p is an issue. But if the same channel is streaming on youtube I can stream at 1080p no issues.
Click to expand...

Youtube has a local CDN, I can stream twtich 1080p with livestreamer but its a pain, If I use a local VPN to a VPS I setup I can stream twtich at 1080p no problem this tells me its a coolidea's issue.
 
nightshade21099 said:
If I use a local VPN to a VPS I setup I can stream twtich at 1080p no problem this tells me its a coolidea's issue.
Click to expand...

Not really, if your VPS is natting your traffic which it would have to it would manage the TCP scaling for you, which is why it works better. Kind of like a proxy would, what OS is running on the VPS?
 
DJZassie said:
You are correct. The direct connect was terrible so you set me up to go via your peer in London.
That worked great until now where it seems to be very similar to how it was for the direct connect.

But i'm 90% sure this isn't an issue with AWS as i have had mates on different ISP's try and use my AWS Instance and they have not had the same issue as me.
Click to expand...

Well it doesnt make much sense that via their own Direct connect or via another completely different path of our own you have the same performance, the only common factor there is AWS. We can drop our peers in London and run over HE and see if this makes any difference. But I imagine it would just use HEs peer at LINX as well.
 
SharkBait said:
I know it might not help much, but here is mine on Vodacom 100/50mbps.
View attachment 520195
Click to expand...

This is from one of our edges where one of the Vumatel handoffs are:

# ADDRESS LOSS SENT LAST AVG BEST WORST
1 154.0.3.110 0% 3740 0.7ms 0.6 0.4 86.7
154.0.4.50
154.0.4.42
154.0.3.114
2 154.0.4.254 0% 3740 158.5ms 182.1 157.9 372.7
154.0.1.146
154.0.1.106
154.0.1.222
154.0.1.214
3 195.66.226.234 0% 3740 159.5ms 183.7 159.1 486.6
4 137.221.79.33 0% 3740 159.7ms 186.1 159.3 987.1
5 100% 3740 timeout
6 137.221.77.71 0.1% 3739 166.6ms 190.6 165.9 429.5
137.221.77.73
7 137.221.66.35 0.1% 3739 166.6ms 190.5 166 506.1
137.221.66.39
137.221.66.37
8 10.109.1.26 0.1% 3739 166.7ms 189.9 165.8 448.2
10.109.1.38
10.109.1.18
9 185.60.115.40 0.1% 3739 167.7ms 189.3 165.5 446.5


And the one packet loss seems to happen on Blizzards network.
 
PBCool said:
Well it doesnt make much sense that via their own Direct connect or via another completely different path of our own you have the same performance, the only common factor there is AWS. We can drop our peers in London and run over HE and see if this makes any difference. But I imagine it would just use HEs peer at LINX as well.
Click to expand...

I didn't say it was the same...
What i was saying is that the current implementation is slowly deteriorating to the same level as the direct connect.

Weirdly enough, if i run an MRT continuously while my connection is live it seems more stable than when the MRT isn't running. Not sure if thats all in my head, but it seems to be the case haha
 
Last edited:
DJZassie said:
I didn't say it was the same...
What i was saying is that the current implementation is slowly deteriorating to the same level as the direct connect.
Click to expand...

Well nothing is deteriorating on our network :) and we haven't had any major changes.
 
Status
Not open for further replies.
Back
Top
Sign up to the MyBroadband newsletter
X