New Contributor
•
5 Messages
Latency Spikes
75mbps business HSI customer
Setup is: coax -> ARRIS SURFboard SB6183 -> wired ethernet to laptop
Problem is: latency to four IPs (8.8.8.8, 8.8.4.4, 208.67.222.222 and 208.67.220.220) consistently spikes from +-20ms to +-1500ms and then times out. The latency and timeouts occur at the same time to all four IP addresses. My office infrastructure running on a Time Warner circuit shows no latency or spikes at the same time.
I am running both ping and mtr. MTR shows packets being dropped at what I guess is a peering partner (level 3).
I have swapped out the coax cable and the ethernet cable to no avail.
I really don't have the time to call, be put on hold for 30+, have the same conversation again and again and how to "reboot the modem" and "do a speed test".
Could a Comcast employee on here give me some indication on how to get this resolved? This is not business class service. It's not even resedential level service with the frequency it's occurring.
Thanks
kmaxx
New problem solver
•
20 Messages
9 years ago
Yeah, there is something going on here. There's been a steady decline the past couple of months in regards to speed/latency. I'm near the Bay Area and there seems to be some really bad routing/peering and it's causing the service speed to drop to a lot of sites, can't seem to get any answers. The phone support used to be great but it's like they're trying to combine the residental and business support into the same support model.
0
0
bizinternetcustomer
New Contributor
•
5 Messages
9 years ago
Phone support is abysmal.
Since I have my "own modem" that "they cannot get into" - they will need to send a tech out. I offered them access to all the salient data, I explained exactly what I was doing to generate the stats on packet loss and latency but it was like I was talking Greek to them.
I asked for an escalation and they insisted they'd need to roll a tech.
Comcast, if you're reading this: start treating more technically inclined customers less like sheep and more like, well, technically inclined customers. If we are telling you that mtr is consistently showing packet loss and high latency between our device and your IP router, then it would behoove you to run some tests from the router back to our devices.
I've had to cancel a business HSI line once before and they sent a debt collector after me. It wasn't until I'd threatened to take the stats and RRD graphs to the tech blogs that they backed down. Unfortunately when there are no other providers... You're screwed.
0
0
bizinternetcustomer
New Contributor
•
5 Messages
9 years ago
To be fair: I emailed the account manager who set us up with our HSI business connection last night (Weds) and just a moment ago (Thurs am) her supervisor called me back and gave me a status update that the tech support team has been working on an upstream issue since 4am eastern and he'd get back to me with an ETA when one was available.
Now it's been escalated I am 99% confident the issue will be resolved: the 2nd and 3rd line techs I've worked with in the past at Comcast are nothing but thorough and know exactly what needs patching/rerouting and where.
Will report back if this is resolved but for now, my recommendation would be to hit your account manager and ask for some support with escalation. Avoid phone support at all costs.
0
0
kmaxx
New problem solver
•
20 Messages
9 years ago
Extreme issue with youtube slowness every evening past 5PM, something is going on in the bay area during this time:
Bad route through sunnyvale doing a youtube.com tracert.
0
0
xz4gb8
Problem solver
•
117 Messages
9 years ago
kmaxx,
The trace output you posted does not indicate a bad route through Sunnyvale. If that were the case, the trace would end at that point.
The result at
simply indicates that the router is too busy to respond to ICMP ECHO requests as used by tracert, traceroute, ping, and the like. The reply time of 13 ms several hops further is a realistic estimate of the round trip time to several hops past what you called a "bad route", indicating packets traverse intermediate nodes with no problems. That is, all the routers are doing routing and doing it apparently well.
What this means is that your youtube problem is not disclosed by this tracert data. You might some useful data pinging the youtube server directly. Most probably there is a particular network link which becomes overloaded at that time of day. It could be your local Comcast node or possibly an interconnect between Comcast and a youtube content source. The comparison to Time Warner that bizinternetcustomer made leads to the same posibillities.
0
0
kmaxx
New problem solver
•
20 Messages
9 years ago
Correct, the tracert is not a real good test, but just sample of something seeing "off" here. As comparison, here is the same tracert after the usual hours of slowness:
Would this at all indicate that there is something going on with the router in sunnyvale since after the usual evening slowness everything resumes back to normal? I'm talking not only about the slow loading of videos, but even slow loading of the main youtube.com site from my connection.
Edit: I guess this doesn't really show an issue since this router may just be placing the ICMP traffic at low priority, but it's curious as to why this one seems to be responding this way and the others are all ok.
0
0
bizinternetcustomer
New Contributor
•
5 Messages
9 years ago
No progress to report since I was called the other day.
MTR showing that there's consistently a problem between my modem and the next hop - Comcast's router.
Avg response time for the last 24 hours is 600ms, best is 5.1ms, worst is 20085ms. Packet loss is about 2% which isn't terrible but the latency spikes basically muffle the packet loss
The rest of the path is by and large error free relative to this probem hop.
Not that impressed. Impossible to actually run a business using the business class service I am receiving.
0
0
CCMichael
Gold Problem solver
•
575 Messages
9 years ago
Please check your inbox concerning this issue.
0
0
kmaxx
New problem solver
•
20 Messages
9 years ago
Is this showing a problem? Just reran this today to the host with the youtube video r12.sn-ab5l6m7z.googlevideo.com:
My traceroute [v0.85] server707 (0.0.0.0) Wed Feb 3 19:55:02 2016 Keys: Help Display mode Restart statistics Order of fields quit Packets Pings Host Loss% Snt Last Avg Best Wrst StDev 1. 192.168.1.1 0.0% 301 0.2 0.2 0.1 0.3 0.0 2. [redacted] 3. 96.120.90.29 0.0% 301 8.8 9.1 6.5 56.6 4.4 4. te-0-7-0-3-sur03.fairfield.ca.sfba.comcast.net 0.0% 301 9.6 8.9 3.4 49.3 4.1 5. be-203-rar01.rohnertpr.ca.sfba.comcast.net 0.0% 301 19.5 11.2 9.4 76.9 5.4 6. hu-0-18-0-3-ar01.santaclara.ca.sfba.comcast.net 0.0% 301 22.3 14.3 8.0 55.1 5.2 7. be-33651-cr01.sunnyvale.ca.ibone.comcast.net 33.0% 300 33169 29748 26223 33169 2195. 8. be-11624-cr02.dallas.tx.ibone.comcast.net 0.0% 300 13.9 14.7 12.4 64.4 5.7 9. ix-1-3-0-0.tcore1.PDI-Palo-Alto.as6453.net 0.3% 300 25.0 18.4 9.5 64.1 7.0 ix-2-3-0-0.tcore1.PDI-Palo-Alto.as6453.net ix-0-1-1-0.tcore1.PDI-Palo-Alto.as6453.net 10. if-2-2.tcore2.PDI-Palo-Alto.as6453.net 3.3% 300 92.8 93.4 15.7 143.8 10.7 11. if-5-2.tcore2.SQN-San-Jose.as6453.net 34.7% 300 95.5 93.6 14.5 129.6 11.7 12. if-1-2.tcore1.SQN-San-Jose.as6453.net 13.7% 300 93.1 93.9 17.1 120.9 8.9 if-29-2.tcore1.CT8-Chicago.as6453.net 13. if-12-2.tcore1.NTO-New-York.as6453.net 91.0% 300 93.6 97.5 93.5 98.6 1.2 if-26-2.tcore2.NTO-New-York.as6453.net 14. if-8-2.tcore2.NTO-New-York.as6453.net 2.3% 300 94.9 95.4 89.5 139.7 5.1 216.6.81.46 63.243.128.6 15. 216.6.81.46 2.0% 300 93.2 95.5 89.6 184.6 7.3 cache.google.com 16. cache.google.com 1.7% 300 94.9 95.2 88.3 126.4 4.00
0