Brutal latency/ping Recently

[t27c]

I am wired, with the gigabit package and all of the sudden have gotten constant ping spikes for over the last few days. I haven't been able to play any games online because the crazy ping spikes and latency make it completely unplayable. My speeds are what they are expected to be, no issues there. I have tried hard wiring straight into the modem but alas, the issue still persists. I have tried switching cables, power cycling my devices, factory resetting my devices. The issue still persists. I have called and contacted Rogers multiple times and they say everything seems fine on their end. But still, the issue persists and is steady. resulting in me not able to use any of my gaming devices due to the brutal and constant ping spikes. It's frustrating paying over $100 a month for internet I cant use for the things I want it for. Any help or suggestions are welcomed and appreciated. Thank you

*** Edited Labels ***

[BigJay1]

In Brampton as well....Bovaird/Torbram, as I reported earlier on, since last week, I am not so much basing my results on pings as I am not a game player....BUT, The wild spikes in speed is crazy....950 Mbps one second, 10 the next...

Buffering issues with Youtube, watching CNN live, whatever...its very very unstable.  I actually have Bell Fibe TV, and with that comes free internet now.....I am actually using that for some of my devices until Rogers gets their act together on this one.  The SECOND FTTH is available in my area...."nowhere in sight unfortunately"  I am outta here..

[sbenninger]

My kids have informed me that no lag spikes in Overwatch or other online games today so that is a positive. My smokepings are very clean as well.

My DSLReports speed tests are down in speed ~650/27 but A's across the board for a change. Had trouble completing the tests since the 23rd. Before the issues started i was ~940/29.5 (QOS to prevent buffer bloat) and A+. Rogers Speedtest also down in speed to the same as the DSLreports test. 

For now I will take the lower speed if it means the connection is stable. Rogers is either going through upgrades or having capacity issues i would think.

[Mikec_81]

I just had another spike just now clearing some heroic visions in World of Warcraft.  Erindale area in Mississauga.  Ship has already sailed for me, ordering another service.  We'll see how tonight's raid goes.

[mrdiu]

I just checked this out, I got an F for BufferBloat and A for Quality.

[sbenninger]

@mrdiu Did you get your full speed that you are supposed to get? If you are using the rogers supplied gateway and getting you full upload speed you will always have bufferbloat as once you max your upload out your latency will also increase due to the nature of asynchronous communications. Using a router that you can limit the upload to a value slightly lower than max upload (in my case 29.5Mbps as I am *supposed* to get 30Mbps) will improve your buffer bloat greatly.

[herrshaun]

@sbenninger And if "just get a router" isn't an option?

[BigJay1]

One interesting point I have noticed over the last few days, I have not seen it mentioned.....When doing a speedtest, I always of course select the Rogers server........As mentioned, speeds are all over the map.....However, when I select the "Bell Mobility" server for the test, initially at least, the speeds always start off much faster, then degrade.  Tells me there is definitely something going on with the Rogers side of things.  Another point, With all the people working at Rogers...techs, staff, engineers etc....Scattered all across the GTA, I would assume they are all experiencing the same issues...Are Network operations staff immune to this at home?  I find it hard to believe that no one internally working for Rogers has not brought this up within the company itself.....How could a network engineer, at home, using rogers, not notice this and want to investigate???????

Jason

[sbenninger]

@herrshaun I am in no way suggesting the fix the issues outlined in this thread is to get a new router but merely describing that you will not be able to remedy bufferbloat in general with rogers supplied gateways.

I am running my own high end router (can push upwards of 10Gbps) and I have still experienced the same ping spikes and latency as everyone else here has described over the last few weeks. Those issues are only issues that Rogers can fix.

[Swifty00]

sbenniger - the problem has nothing to do with hitting the upload or download throughput limits.

The "bufferbloat" is not being caused by throughput restrictions.  If that was the problem there wouldn't be any problem with gaming packets, or ping tests for that matter.  I run a ping test on the gateway server and I get >500% jitter every time.  No throughput problems at all.

The problem is that the routers can't process the policing and shaping queues fast enough.  It probably would be better called "Queuebloat", but the net effect is still that the buffer is getting packed with millions of small packets awaiting policing one by one, not large downloads that have been cleared to go.

This is why the Rogers test works fine and the Keystone cops at Rogers are blundering around without a clue.  They are obsessed with throughput.

They have masses of throughput, but they have totally underpowered routers running it.  Their network code regulating traffic has probably become so bloated that the routers cant process packets and get traffic onto the network fast enough.

I suspect they are sitting their scratching their heads wondering why the network is at 50% capacity, but they are still having this problem.

[sbenninger]

@Swifty00 I fully understand what bufferbloat is but it becomes greatly exaggerated when you reach your capped upload speed with an async connection like cable docsis. When you reach your hard cap then packets get rate limited (dropped in most cases) and you need to wait for queues to flush, fill, rate limit then de-queue again etc. This will also affect your download at the same time due to it having to wait every cycle for the above to complete. 

Ever notice on a speedtest the upload is never a constant speed..it alway fluctuates above then below your rated speed. That is the affect while at the same time causing massive latency spikes (showing as bufferbloat). All i am saying is that if you prevent hitting the hard cap with nicer qos style queues you will reduce the bloat in these tests. Run a continuous ping to a fast site like google dns...then start a large upload that can saturate your upload speed and watch what happens.

Again this has nothing to do with the current issues going on . They may have poorly configured or underspec'd routers causing the spikes but not related to what I am specifically talking about which I admit is off topic so will leave it at that.

[Swifty00]

Yes that can happen.  I am almost certain that is what Rogers think the problem is as well.  As I said, they are obsessed with throughput capacity.  I'm sorry I jumped at you.  I am so frustrated with this.

The Rogers test even has a jitter marker on it and is obviously just measuring jitter on a throughput test, which is obviously fine.  The tech then looks at me proudly, as if he has proved there really isn't any jitter on the network.

Then I show him my ping test to the Gateway address and he looks blank.   Anyway, I have another tech coming out at noon tomorrow, hopefully this one will have a clue, but I'm not holding out any hope.

It would be comical if it wasn't so irritating.  They did some work on our local nodes to day speeding up the branch nodes.  The problem got worse.  Guess why?  Now the outer nodes are running faster the edge router buffer/queue is filling up faster and the jitter gets worse.  Sigh!

[BoldAdanac1]

Important Update: 

I lost internet for an hour earlier this morning. When it came back all problems were gone! Also clearly means rogers did something even though upon calling they said there were no changed to any of my tickets. Also I was banned from the forum so I assume all my old posts are gone. Essentially I have been having these issues since the 14th of February. After 4 in house visits, and two maintenance team visits to the varies poles around my area, the issue was left unsolved and escalated to who ever is higher than maintenance (I assume engineers). I received a call from one of the actual rogers maintenance guys who I have known for some time now and as of this morning he hadn't heard any updates on the case. But then as said above my internet went out for approx 1HR and when it came back all issues were magically gone. 

[geokilla007]

I've signed up for Bell FTTH. On to greener pastures if Bell does the install the way I want them to...

[Datalink]

@BoldAdanac1 what modem are you using?  Can you post the signal levels please, when you have time? That might provide some hints as to what's transpired. 

[Datalink]

@BoldAdanac2 Thanks very much.  I was looking for any evidence of multiple DOCSIS 3.1 OFDM downstream channels running or perhaps any upstream DOCSIS 3.1 OFDM channels running, but, nope, not at all.  So, the signal types look like they usually do.  

Fwiw, your downstream DOCSIS 3.0 channels (1 to 32) are a little high as are the upstream.  Thats an unusual combination.  They're all well within spec so you shouldn't run into any issues.  Just an observation.  Is there a powered amplifier installed on the inbound cable from the street??  Just wondering.

[BoldAdanac2]

@Datalink I know the maintenance team installed something new on the pole right outside my house but I am not sure what it was. I was also given a filter from one of the last techs as the signals were even higher originally .

[Datalink]

If your cable runs overhead from the utility pole, it sounds like the techs may have replaced the local tap which connects you and your immediate neighbours to the Rogers cable system.  They look something similar to this:

http://www.acmetronix.com/at/index.php?m=content&c=index&a=show&catid=15&id=53

https://www.multicominc.com/product/multicom-mtar-pi-x-x-plug-in-tap-series/

That local tap is either mounted on the utility pole, or in the case of underground cabling, its located within a nearby pedestal. 

[sbenninger]

Well all was great with my connection yesterday for the first time since the 23rd...then 2am (the Rogers witching hour) rollls around and something changed and back to latency spikes and packet loss. They are obviously working on fixes but what is anyones guess.

Would be nice for some sort of communication from Rogers on the issues but I am guessing they are trying to avoid bad press.

[Swifty00]

Run the DSL Reports test:

http://www.dslreports.com/speedtest

It will flood the Network with small packets and test bufferbloat.  Even at off-peak times when you can just about game and spikes are not too bad, this report will test your connection and you will find out if they really have fixed it.

[Swifty00]

PS - Run this test while your kids are playing Overwatch and listen to them howl - if you really want to have fun!

[Flameshow]

I can't play anything at all, in league of legends it spikes to 400ms for 2-3 seconds at the random, Apex Legends became a running in place simulator, Overwatch is a nightmare, no shots register at all... and call of duty is just purely unplayable, i get killed before I even see the enemy. What can I say, small Indie company that doesn't have the resources to solve the matter...... Highly considering switching to Bell but I am afraid the same problem might reoccur....