Fireblade
join:2008-08-27
St Catharines, ON
 ·Cogeco Cable
 ·Vonage
| nLayer Sucking Hard
This has been happening for weeks, can't play on any L4D servers outside New York and maintain a ping lower than 130ms. I thought Cogeco used peer1, not this nLayer junk?
--
I love fish sticks. I love putting fish sticks in my mouth. |
|
exseven
Premium,VIP
join:2003-05-23
Beamsville, ON
3 edits | so bell.ca = nLayer now? Cogeco has multiple peers, not just peer1 |
|
Fireblade
join:2008-08-27
St Catharines, ON
·Cogeco Cable
·Vonage
| 15 247 ms 200 ms 204 ms as20473.tge5-2-106.ar1.dfw1.us.nlayer.net
14 91 ms 231 ms 199 ms as20473.tge7-2-1615.ar2.ord1.us.nlayer.net
Is Bell?
--
I love fish sticks. I love putting fish sticks in my mouth. |
|
exseven
Premium,VIP
join:2003-05-23
Beamsville, ON
1 edit | reply to Fireblade
said by Fireblade  :This has been happening for weeks, can't play on any L4D servers outside New York and maintain a ping lower than 130ms. I thought Cogeco used peer1, not this nLayer junk? i added the bolding, look up on your traceroutes and you will see cogeco hands off to bell in that trace, not nLayer |
|
Fireblade
join:2008-08-27
St Catharines, ON
1 edit | reply to Fireblade
Re: nLayer Sucking Hard
I take it back.
P.S. Bell still sucks though.
--
I love fish sticks. I love putting fish sticks in my mouth. |
|
Richard S
@comcast.net
| reply to Fireblade
Hello (due to the power of google search alerts :P),
I think you're confused about how traceroute works. Just because there is a spike in the middle doesn't mean the latency is actually high for normal traffic, and as you can see the latency drops back down to normal in future hops. This is an artifact caused by a very busy router CPU on that one hop (it's a Cisco software bug which causes high CPU utilization, we're currently planning a maintenance to upgrade the router, but there is no actual impact it is purely cosmetic). There are also pending upgrades with Bell which will provide a more direct route for their users in Toronto trying to reach Chicago, give it another week or so then check back.
For more information about how to read traceroute correctly, and why the high latency hop you pointed out isn't actually impacting traffic, please take a look at:
»www.nanog.org/meetings/nanog47/p···_Sun.pdf |
|
humble226
join:2009-10-26
Chicago, IL
| reply to Fireblade
Pending that additional interconnection with Bell (it was actually due in today but got pushed back a few days), I've made some routing changes for you which optimize the return path from those Chicago servers back to Cogeco so you don't have to go all the way out to New York, Montreal, etc. This should get you down from 51ms RTT to about 30ms. After the new circuits are in, you should be down to around 10-15ms to Chicago (depending on which fiber path Bell takes, the short one via Detroit or the long one around Lake Erie). That's about the best we can do right now, if anyone was handing this off to us in Toronto we'd have more control over it but alas the inbound side is more difficult to influence.
In the future, feel free to drop us an email at noc@nlayer.net if you see a route you think is bad. If there is a suboptimal routing issue we really do want to know about it so we can fix it, but worst case we'll at least be able to explain what the problem is.  |
|
Fireblade
join:2008-08-27
St Catharines, ON
·Cogeco Cable
·Vonage
| said by humble226 :Pending that additional interconnection with Bell (it was actually due in today but got pushed back a few days), I've made some routing changes for you which optimize the return path from those Chicago servers back to Cogeco so you don't have to go all the way out to New York, Montreal, etc. This should get you down from 51ms RTT to about 30ms. After the new circuits are in, you should be down to around 10-15ms to Chicago (depending on which fiber path Bell takes, the short one via Detroit or the long one around Lake Erie). That's about the best we can do right now, if anyone was handing this off to us in Toronto we'd have more control over it but alas the inbound side is more difficult to influence. In the future, feel free to drop us an email at noc@nlayer.net if you see a route you think is bad. If there is a suboptimal routing issue we really do want to know about it so we can fix it, but worst case we'll at least be able to explain what the problem is. Wow, thanks I'm actually very impressed at the quality of service you've provided. Most companies would turn a blind eye, I take back the nLayer junk comment. What you did improved my ping to 8.12.21.106 by 60ms, from 150 to 90 and it's stable without any spikes. Sure 90 isn't ground breaking but for a $6 USD / month L4D server I'm not complaining.
--
I love fish sticks. I love putting fish sticks in my mouth. |
|
Fireblade
join:2008-08-27
St Catharines, ON
·Cogeco Cable
·Vonage
| reply to Fireblade
Anyway to improve the ping to this Dallas server? Our clan is trying to centralize our latency because we're all from different parts of the country. So far, this Dallas server using Bell/Internap is by far offering superior latency vs. Bell/nLayer -- if improvements can be made I would appreciate it, if not no worries. Also, L4D servers run at a tick rate of 30, so ping in game will be significantly higher than lets say a CS server running a 66/100 rate.
--
I love fish sticks. I love putting fish sticks in my mouth. |
|
nevertheless
Premium,VIP
join:2002-03-08
Burlington, ON
·Cogeco Cable
| reply to Richard S
An interesting factoid from this link:
quote:
• A round-trip around the world at the equator, via a perfectly straight fiber route, would take ~400ms due solely to speed of light propagation delays speed-of-light delays.
--
Some people think I'm an idiot. I disagree, but idiocy is subjective--so they may well be right. With this in mind, take everything I post with a grain of salt, eh? |
|
humble226
join:2009-10-26
Chicago, IL
| reply to Fireblade
said by Fireblade :Anyway to improve the ping to this Dallas server? Our clan is trying to centralize our latency because we're all from different parts of the country. So far, this Dallas server using Bell/Internap is by far offering superior latency vs. Bell/nLayer -- if improvements can be made I would appreciate it, if not no worries. Also, L4D servers run at a tick rate of 30, so ping in game will be significantly higher than lets say a CS server running a 66/100 rate. Same problem, Traffic is going Toronto - Montreal - New York - Washington DC - Atlanta - Dallas and back again, instead of the much shorter and more direct Toronto - Chicago - Dallas. Applied the same fix, routing outbound traffic via a more direct route, bringing the round-trip latency down from 77ms to 53ms. Same as with Chicago, this is a temporary kludge while waiting for the activation of the more direct route. When finished it should be around 37ms from Dallas and look something like this:
ras@re1.cr1.dfw1> traceroute ar1.tor1.nlayer.net
traceroute to ar1.tor1.nlayer.net (69.31.143.246), 30 hops max, 40 byte packets
1 xe-2-2-0.cr2.ord1.us.nlayer.net (69.22.142.2) 21.611 ms 21.621 ms 21.654 ms
MPLS Label=764347 CoS=0 TTL=1 S=1
2 ae3.cr1.ord1.us.nlayer.net (69.31.111.153) 21.681 ms 21.650 ms 21.606 ms
MPLS Label=324926 CoS=0 TTL=1 S=1
3 tge1-4.ar1.tor1.ca.nlayer.net (69.22.142.78) 37.191 ms 37.211 ms 37.900 ms |
|
Fireblade
join:2008-08-27
St Catharines, ON | Do you know the approximate date for that transition process?
--
I love fish sticks. I love putting fish sticks in my mouth. |
|
humble226
join:2009-10-26
Chicago, IL
| said by Fireblade :Do you know the approximate date for that transition process? Here you go:
|
|
Fireblade
join:2008-08-27
St Catharines, ON | reply to Fireblade
PERFECT, that server I was pinging 117, is now 66 constant in-game. Very good, I'm pinging just as good as my friends in Phoenix. Nice job.
--
I love fish sticks. I love putting fish sticks in my mouth. |
|
