congestion between Cogent and CenturyLink
Hello, We send periodic 10-15Mbps bursts of traffic to a business partner and it appears to transition from Cogent to Century Link in Atlanta. During the day performance is normal and latency appears acceptable on a trace route. 12 ms 13 ms 12 ms te0-6-1-7.rcr21.msp01.atlas.cogentco.com [38.88.188.41] 26 ms 26 ms 27 ms be2410.ccr22.ord01.atlas.cogentco.com [154.54.7.229] 43 ms 44 ms 44 ms be2099.ccr22.atl01.atlas.cogentco.com [154.54.28.74] 44 ms 44 ms 44 ms be2051.ccr21.atl02.atlas.cogentco.com [154.54.0.162] 43 ms 42 ms 43 ms qwest.atl02.atlas.cogentco.com [154.54.13.30] 49 ms 49 ms 50 ms min-edge-10.inet.qwest.net [205.171.128.154] But after hours latency spikes and throughput drops to less than 1Mbps. te8-3.ccr01.msp01.atlas.cogentco.com (38.88.188.41) 4.603 ms te0-0-0-19.mpd22.ord01.atlas.cogentco.com (154.54.1.230) 17.838 ms be2099.ccr22.atl01.atlas.cogentco.com (154.54.28.74) 33.156 ms be2051.ccr21.atl02.atlas.cogentco.com (154.54.0.162) 36.449 ms qwest.atl02.atlas.cogentco.com (154.54.13.30) 89.583 ms min-edge-10.inet.qwest.net (205.171.128.150) 89.806 ms Century Link stated that Cogent is oversubscribing the link, and that they've requested Cogent resolve the problem, but that action has yet to be taken. I've tried reaching out to Cogent but as we're not a direct customer they wouldn't provide assistance. Has anyone else seen similar issues? Thanks, Aidan
With cogent? Now you will be asking us if the Pope is really Catholic :) On 28-Feb-2014 7:43 AM, "Aidan Scheller" <aidan@aodhandigital.com> wrote:
Hello,
We send periodic 10-15Mbps bursts of traffic to a business partner and it appears to transition from Cogent to Century Link in Atlanta. During the day performance is normal and latency appears acceptable on a trace route.
12 ms 13 ms 12 ms te0-6-1-7.rcr21.msp01.atlas.cogentco.com [38.88.188.41]
26 ms 26 ms 27 ms be2410.ccr22.ord01.atlas.cogentco.com [154.54.7.229]
43 ms 44 ms 44 ms be2099.ccr22.atl01.atlas.cogentco.com [154.54.28.74]
44 ms 44 ms 44 ms be2051.ccr21.atl02.atlas.cogentco.com [154.54.0.162]
43 ms 42 ms 43 ms qwest.atl02.atlas.cogentco.com [154.54.13.30]
49 ms 49 ms 50 ms min-edge-10.inet.qwest.net [205.171.128.154]
But after hours latency spikes and throughput drops to less than 1Mbps.
te8-3.ccr01.msp01.atlas.cogentco.com (38.88.188.41) 4.603 ms
te0-0-0-19.mpd22.ord01.atlas.cogentco.com (154.54.1.230) 17.838 ms
be2099.ccr22.atl01.atlas.cogentco.com (154.54.28.74) 33.156 ms
be2051.ccr21.atl02.atlas.cogentco.com (154.54.0.162) 36.449 ms
qwest.atl02.atlas.cogentco.com (154.54.13.30) 89.583 ms
min-edge-10.inet.qwest.net (205.171.128.150) 89.806 ms
Century Link stated that Cogent is oversubscribing the link, and that they've requested Cogent resolve the problem, but that action has yet to be taken. I've tried reaching out to Cogent but as we're not a direct customer they wouldn't provide assistance.
Has anyone else seen similar issues?
Thanks,
Aidan
+1, which semi-large eyeball does Cogent NOT have capacity problems to? On 2/28/2014 午前 11:55, Suresh Ramasubramanian wrote:
With cogent? Now you will be asking us if the Pope is really Catholic :) On 28-Feb-2014 7:43 AM, "Aidan Scheller" <aidan@aodhandigital.com> wrote:
Hello,
We send periodic 10-15Mbps bursts of traffic to a business partner and it appears to transition from Cogent to Century Link in Atlanta. During the day performance is normal and latency appears acceptable on a trace route.
12 ms 13 ms 12 ms te0-6-1-7.rcr21.msp01.atlas.cogentco.com [38.88.188.41]
26 ms 26 ms 27 ms be2410.ccr22.ord01.atlas.cogentco.com [154.54.7.229]
43 ms 44 ms 44 ms be2099.ccr22.atl01.atlas.cogentco.com [154.54.28.74]
44 ms 44 ms 44 ms be2051.ccr21.atl02.atlas.cogentco.com [154.54.0.162]
43 ms 42 ms 43 ms qwest.atl02.atlas.cogentco.com [154.54.13.30]
49 ms 49 ms 50 ms min-edge-10.inet.qwest.net [205.171.128.154]
But after hours latency spikes and throughput drops to less than 1Mbps.
te8-3.ccr01.msp01.atlas.cogentco.com (38.88.188.41) 4.603 ms
te0-0-0-19.mpd22.ord01.atlas.cogentco.com (154.54.1.230) 17.838 ms
be2099.ccr22.atl01.atlas.cogentco.com (154.54.28.74) 33.156 ms
be2051.ccr21.atl02.atlas.cogentco.com (154.54.0.162) 36.449 ms
qwest.atl02.atlas.cogentco.com (154.54.13.30) 89.583 ms
min-edge-10.inet.qwest.net (205.171.128.150) 89.806 ms
Century Link stated that Cogent is oversubscribing the link, and that they've requested Cogent resolve the problem, but that action has yet to be taken. I've tried reaching out to Cogent but as we're not a direct customer they wouldn't provide assistance.
Has anyone else seen similar issues?
Thanks,
Aidan
I'm already seeing a huge improvement to Comcast after Netflix moved a lot of traffic off of the ports. On Feb 27, 2014, at 22:21, Stephen Frost <sfrost@snowman.net> wrote:
* Paul S. (contact@winterei.se) wrote:
+1, which semi-large eyeball does Cogent NOT have capacity problems to?
Soon, Comcast... Given what's going on w/ them and Netflix.
Thanks,
Stephen
I saw the same effect after the Netflix peering started. http://imgur.com/a/aVFAS On Thu, Feb 27, 2014 at 10:22 PM, Jason Canady <jason@unlimitednet.us>wrote:
I'm already seeing a huge improvement to Comcast after Netflix moved a lot of traffic off of the ports.
On Feb 27, 2014, at 22:21, Stephen Frost <sfrost@snowman.net> wrote:
* Paul S. (contact@winterei.se) wrote:
+1, which semi-large eyeball does Cogent NOT have capacity problems to?
Soon, Comcast... Given what's going on w/ them and Netflix.
Thanks,
Stephen
participants (6)
-
Aidan Scheller -
Edward Roels -
Jason Canady -
Paul S. -
Stephen Frost -
Suresh Ramasubramanian