Connectivity
Connectivity
Didn't find what you were looking for?
We have advanced search options to make it easier to locate posts, questions and answers on this community.
More information can be found at Advanced Search Options
If you are looking for something specific, please check if someone else has already asked or answered the same question.
Any suggestions why my upload latency is so horrible?
When Spectrum ran fiber out to our area last year, it seemed to work great. Just as it did when we had Spectrum in San Antonio, TX and Seguin, TX. But not long after the good times, came the bad times. It has gotten worse over time also. At first it was just a lag spike here or there, then it was so common I quit trying to play online games. My kids still try, but do nothing but complain because of the lag. I work from home also and have just had to deal with 1 bar of Spectrum wireless signal to handle work calls, as using calls over wifi has become unbearable for myself and the person on the other end of the call. I've attached a copy several days of results from Speedtest CLI that I had run every hour if the PC was on. I tested multiple servers before running the automated test and Frontier always came out with the best results, so I locked the test to that single server for testing purposes.
Any idea why the upload latency is just horrific?
Best Answer
-
After several calls and email and online chat messages over the last 18 hours, I was finally able to get my speed package upgraded. I'm having much better results when using Voip and the tests I have ran have been much improved! I can only assume the upload latency issues I was experiencing were due to a bottleneck (or congestion?) when under load with only 10mbs upload speed on the previous package. This test was one of the highest in terms of latency thus far after the upgraded speeds and I'm not going to complain one bit.
https://www.waveform.com/tools/bufferbloat?test-id=cd723ef4-3116-4364-8008-c009c0c8fded
Thanks for all of the help and suggestions from everyone. :)
0
Answers
-
Hello and welcome to the Spectrum Community!
I do apologize for the issues with our internet service. I was able to find your account using your community information. I am not seeing any issues with your modem or the service. Please try the troubleshooting steps here to see if that helps.
If you still have issues you may need to have a technician come out and take a look at the issues for you.
-Tyleen
0 -
I’ve went over those steps numerous times and unfortunately it hasn’t made a difference. I have updated my router firmware each time Asus releases an update over the past year, I’ve restarted the spectrum equipment and my router countless times and I’ve ran test with WiFi off, ensuring my PC was the only device connected to the internet (all test are done via wired connection to rule out WiFi signal issues) and nothing has changed the upload latency. A tech came out previously and could not find the cause on site and they changed out the equipment once, thinking that may be it. He even ran test with my PC connected directly to the spectrum equipment, bypassing the router, (which I have done on a couple occasions since then) and the upload latency remains.
my neighbor across the street has latency issues as well, but it doesn’t bothering him as much since they mainly use it for watching videos or just visiting websites. Streaming videos is great for us too, but when it comes to first person games or voip it’s like I’m back on dialup in 1999.0 -
@JasonC79 Have you tried a different ethernet cable between the modem and router, or between the router and your computer? It looks like the last technician was quite a while ago, if different ethernet cables don't help I'd recommend scheduling another appointment.
0 -
alright, I have replaced the cat 6 cables going from the spectrum equipment to my router and from my router to my PC. Unfortunately, our service is out again so I’ll report test results when service gets restored.
0 -
still having the same upload issues after swapping out with all new cables.
0 -
Aside from possible signal quality issues, there could be a fair amount of buffer bloat in play.
The Asus router should have a QoS section where you can enable Cake QoS to mitigate a good bit of it.
Just the most basic settings can make a noticeable difference in many cases... don't have to get too complicated with it.
While you can get advanced and set transfer caps per device or by packet types and what not... it works well just keeping it simple by setting the universal boundaries for all transfers.
Just set the basics... expected upload and download bandwidths.
It will try to balance the latency to keep it more consistent by controlling flow to prevent the off and on filling and flushing of the buffers (basically, all your transfers normally go in bursts with pauses in between). The nature of how it works will reduce your peak transfer rates reported on some speed tests… but it should keep the flow and average rate more consistent. The impact will vary based on where you set them between your typical transfer rates and your subscribed tier.
You can run before/after tests and learn more about buffer bloating at the waveform site:
https://www.waveform.com/tools/bufferbloat
0 -
alright, I have replaced the firmware with the Asuswrt-Merlin build, enabled CAKE, and here are my before and after results from waveform.
Before replacing firmware and running standard adaptive QoS:
https://www.waveform.com/tools/bufferbloat?test-id=2b21cfad-56e7-4bd9-90be-680b46bb504d
After replacing firmware and enabling CAKE:https://www.waveform.com/tools/bufferbloat?test-id=84db475f-8bbb-4ff0-9e72-9675d757f333
If there are specific changes I need to make to really make the firmware and CAKE implementation more effective and please feel free to let me know.0 -
Ooof... you definately have an upstream signal problem. Depending on a few factors, you should be consistently getting at least 15-22mbps on the upstream, even if it is still DOCSIS 3.0 in play.
It may be only latching 1 usable 3.0 channel (which drastically reduces potential bandwidth), and the modem may be screaming to try to reach the headend because of the noise, which is also getting a lot of errors and just adding insult to injury.
It is possible (and hopeful) you have some loose/corroded connections or some compromised shielding or bad ground somewhere causing the return path to get corrupted. Such things are <usually> easily resolved.
Could do a rudimentary inspection of such coax connectors/cables before calling for a service tech visit. If you spot anything sus, be sure to mention it to them during the visit. They routinely replace connectors and splitters when they show signs of degrading... can even arrange to replace SOME cabling, depending where the cable runs.
If you open connections, the copper wire and insides of the connecting faces should be clean, dry, and at least a little bright… not necessarily shiny, but not dull and cloudy. You want them to be slightly better than just finger tight (not torqued up, but clearly tight enough it will take some effort to unscrew them without any tools). The outer shielding should obviously be intact with no splits or rough edges, no crimping, and no bends sharper than about the curve you would get around a 2-liter soda bottle.
If you can see where any outside coax comes into the building, there should be a ground strap/wire of some sort that connects to a junction and goes right into the ground or possibly to another strap or pipe that goes into the ground. This may be just a cable running through the wall into a store room, or it could be a sealed box that you can see the coax run through and then the ground strap might be seen coming out of that box.
That may sound strange with the expectation of fiber optic cable runs, but the system as a whole is actually hybrid... some portions will be fiber, some will be center copper coaxial cable. The coax actually functions as an antenna carrying radio waves through it, so it is important to maintain certain standards to preserve signal integrity inside the "tube".
0 -
As for the cake settings, can usually strike a nice balance setting the limits a little under what you typically get on tests.
It will have a slight clamping effect... so if you were to say put it at 500/20 (base plan in some markets) even though you may routinely see up to 524/22 on tests, you may actually clamp it to around 494/17 or worse depending on the routing in play. Setting it closer to the test, like 520/21, and you may still come out with like 515/19 so it doesn't feel so bad.
But that massive hit on the upstream NEEDS to get flushed out. Last time I saw it go sideways that bad, the landscapers had nicked an exposed portion of my outside drop and stripped off a few inches of the shielding... exposing the center conductor.
In the short term, could try setting the upstream limit to about 10.5, but it may take notching it back to about 9.5 or lower to give it a good chance of dialing back that loaded latency.
Not uncommon to see 8 lines sharing 4 channels in the 3.0 setups, with each channel only offering around 38mbps bandwidth... all shared across 8 households.
Basically, it looks like your line may only have one usable channel in play and is simply not getting much time on the wire to send stuff atm.
0 -
I will inspect what I can access in the morning and report back. Sounds like I’ll need to pay for a service call before too long to help figure out where the issue is. I appreciate everyone assisting with possible solutions thus far though.
0 -
With those kinds of numbers, I would expect no charge for the call.
At least... that has been the norm for residential when there are clear signs something is off like that.
Since you are running your own router though, may want to do a test while connecting a system directly to the modem and capture the results somehow. Just to show it is not tied to your router.
Ruling that out ahead of time can speed things up.
0 -
Another good suggestion. I’ll connect directly to the fiber ONT overnight tonight while everyone is asleep and again first thing in the morning to get those test results with my router removed from the equation.
0 -
I wasn't able to run test last night bypassing the router, but I was able to get the test ran this morning. Below are the results from running two test connected directly to the ONT (router out of the loop) at approximately 7:50 - 7:54 AM CST.
Test 1:
https://www.waveform.com/tools/bufferbloat?test-id=a73f5ca7-1625-4ecf-86c9-328b2038bbb0
Test 2:
https://www.waveform.com/tools/bufferbloat?test-id=4e45227e-a84c-4642-a693-44844c02db0f
After completing the above test, I connected my router back to the ONT and my PC back to the router and ran two test with the router back in the loop. I waited a few minutes to give all of the wireless devices in the house to get back online first.
Test 1:
https://www.waveform.com/tools/bufferbloat?test-id=b4749fd5-81bb-434f-ad5b-e4bf2bc2152e
Test 2:
https://www.waveform.com/tools/bufferbloat?test-id=d6eb3aee-fb4e-462c-b5a2-9a893a20cc7a
0 -
I have a tech scheduled to come out this afternoon, I’ll report back with their findings.
0 -
Tech came out. He changed the ONT, connector on the fiber coming into the house and the power supply on the ONT. He see's there is an issue, but the regional manager here said it was a "me problem", since the ping reports show no issues. The tech sees the issue is when there is a load on the Upload stream, which he acknowledged will not be present on a simple "ping" test. But his boss told him there isn't anything that can be done. I don't believe that, but what do I know.
0 -
Thank you for keeping us posted on this. I am seeing the information you have provided with the work that has been completed. Are you still noticing the concern with the service? -Lyn
0 -
Lyn, yes I’m still having the same issues. I tried taking the techs suggestion and upgrading to the 1Gig plan to see if the additional upload speeds I would receive could help resolve the upload latency, but after completing the upgrade via MyAccount online, I’m still waiting on the speed to be upgraded on the service. It states that I now have the 1 gig package online, but phone tech support and phone sales say they don’t see it. Chat support also said they did not see it active, but said they could complete the upgrade for the $20 additional per month, I agreed for them to complete it. After a moment, they came back to the chat and stated I was all set.
Maybe it’ll be active in the morning. I’ll report back and also post new test results if the upgrade ever gets turned on.
0 -
Below a certain bandwidth usage level, bandwidth limits would have little impact on the latency... and that largely depends on the total load in play and various aspects about how the network operates.
Unless your use case was exceeding the typical capacity for each time slice window granted for each transmit cycle, altering the bandwidth cap would have a very minor impact on the loaded latency in and of it itself.
Large chunks exceeding a current cyclical burst window will get through more efficiently when you change to a plan allowing longer transmit windows... otherwise the smaller chunks like the regular background traffic that a game uses or even VOIP won't really notice much difference from a higher bandwidth cap alone.
It all depends whether your line was aleady getting saturated.
Otherwise it would typically be a high error rate sort of situation (poor signal quality), or something upstream from you getting oversaturated.
Provided You stayed in the realm of a 500-700mbps downstream plan, it is more likely something in the signaling or saturation conditions just happened to improve in the same time frame. Could have been the modem even.
If however you moved to a gigabit plan… there would appear to still be issues that need addressing. Shouldn't be getting clamped that low when things are in order uness some heavy shaping has been set up in QoS or something.
1 -
I agree that there are still likely some issues somewhere. But the tech had the area manager on speaker phone going over things he tried and letting him know there is still an issue. The manager clearly wasn’t concerned or interested in entertaining the idea that they needed to look into it any further. I doubt I can get further than the tech myself, but if you know of a way to escalate the issue up the ladder, I’m all ears.
0 -
Hello @JasonC79;
I reviewed your equipment and it is currently not registering any signal problems. There could be a routing issue that is causing your speeds to drop. Can you run a trace route to a site or server that you are seeing problems with? Post the results here and we can review it to see if there is a potential routing issue.
1 -
Edward,
I ran the tracert below and followed it up with pingplotter after the trace was completed.
Tracing route to www.google.com [2607:f8b0:4023:1004::69]
over a maximum of 30 hops:1 1 ms <1 ms <1 ms syn-2603-808c-0e00-008b-0000-0000-0000-0001.res6.spectrum.com [2603:808c:e00:8b::1]
2 4 ms 4 ms 3 ms 2602-0107-0718-005f-0000-0000-0000-0001.inf6.spectrum.com [2602:107:718:5f::1]
3 2 ms 4 ms 4 ms lag-57.hcr02bnwhtx08.netops.charter.com [2605:6000:0:4::c:236]
4 7 ms 9 ms 9 ms lag-30.mcr11dllbtxlb.netops.charter.com [2605:6000:0:4::2:8016]
5 13 ms 9 ms 9 ms lag-21.rcr01hstqtx02.netops.charter.com [2605:6000:0:4::30]
6 * * * Request timed out.
7 29 ms 28 ms 31 ms 2607:f8b0:825e::1
8 25 ms 25 ms 26 ms 2001:4860:0:1::6ffa
9 34 ms 27 ms 31 ms 2001:4860:0:1::8922
10 22 ms * 23 ms 2001:4860::c:4001:e559
11 21 ms 20 ms 21 ms 2001:4860::c:4003:d02c
12 18 ms 21 ms 20 ms 2001:4860::cc:4001:f891
13 * * * Request timed out.
14 * * * Request timed out.
15 * * * Request timed out.
16 * * * Request timed out.
17 * * * Request timed out.
18 * * * Request timed out.
19 * * * Request timed out.
20 * * * Request timed out.
21 28 ms 30 ms 29 ms rq-in-f105.1e100.net [2607:f8b0:4023:1004::69]Trace complete.
0 -
Hello @JasonC79;
Thank you for the trace routes. The trace routes show normal activity on the network. The time outs and packet loss you are receiving just indicates that the device is not responding to the test. Routing issues can be intermittent. You can run more trace routes to see if it catches an issue. I would also recommend tracing to multiple sites or servers that you have seen issues with. Tracing to a site you are seeing a problem will help identify potential issues with routing to those servers.
1 -
I’ll start logging more trace routes periodically and when I can catch a specific site giving me issues I’ll trace it also. I’ll return here with results that show possible anomalies.
0 -
There is a little wonkiness in the routing for sure... could be tied more to the IPV6 pathing though. The low adoption rates in the US is still a problem.
Ideally, latency should scale more smoothly as distance increases (ie: as you move through the path to the destination). But with fewer nodes out there it reduces options for pathing, so you can get some weird variances.
One of many reasons the gaming industry is still so deeply invested in IPv4... we need to see way better adoption rates for IPv6 to really shine.
So if looking to diagnose an issue for the gamers at home, may want to run tests with the IPv4 switch too... just in case they are following those routes.
It can vary a little depending on the environment, usually something like
tracert -4 example.com
Some iterations of the tool may use / instead of -
1
