Question Ping spikes only on international servers

[silver085]

Hello, basically I confirmed it's happening only on servers that aren't in my country.

I was pinging a lot of servers from a lot of games and the result is always the same. The moment ping spikes appear in foreign server, I then check the log from the 2nd CMD which is doing ping on the server located in my country and that one is always fine.

I'm uploading 2 example pings so you can see, one is German server, 2nd is located in my country - Poland. Both pings ran and were stopped at the same time, so you can scroll to the very bottom and see on the 1st one where the problem appears, and 2nd is fine at the same time.

1.German : https://pastebin.com/96EqBtfD
2.Poland : https://pastebin.com/igsYxPkk

Those pings spikes happen several times a day, after upgrading my internet speed which came with the new router. On my old one, I never had any of those issues for like 10 years.
The new one is : CGA4236TCH1

The PC is connected to my router with ethernet cable and there are no other devices in between.

I was told to do WinMTR tests, and so I did to a MMO server that's also located somewhere in Europe :

So is this some kind of problem with routing, that only my ISP can resolve?

 

[lantis3]

Your ISP will not change route for a personal request.

Pretty much nothing you can do if problem happens somewhere on the internet. Change the route is the only hope that probably can improve the situation.

Use a VPN, probably free Proton VPN to begin with.

[SOLVED] - I have ping issues on Roblox ?

My ping on Roblox is usually around 50-100ms, which is completely usable, but I generally have spikes up to 600+ ms regularly, the only way I can currently fix it is by using a 5G hotspot, but that is a band-aid solution since I will eventually run out of hotspot data. I know there was another...

forums.tomshardware.com

https://www.vpnbook.com/freevpn is another choice.

 

[punkncat]

I don't know if the issue is connected to your problem but over the past couple of days there has been a number of the undersea cables that provide service to both Europe and Africa damaged. There has been a lot of load on other systems as they try to bypass the physical issue with less than optimal routing.

 

[silver085]

@lantis3
I tried both of these VPN's but they just increase in game ping from 80 to 300-400 so everything is just teleporting .

@punkncat
This has been happening since few months, so probably unrelated and besides it's always only me with those issues while playing that MMO. Everyone else that I know from that game can play just fine for hours.

 

[lantis3]

How about Cloudflare Warp?

 

[bill001g]

Your problem is you can only fix what you have control over.

This is mostly your pc, the router and to some small extent the connection you pay your ISP for between your house and the ISP.

The data from both test cases uses the exact same devices so it is not likely there is a issue in these devices. What your traces show is there is some issue far away from your house likely in level 3 network. Do you really think you can just call up level 3 and they will change things for some gamer. Even if you were thinking of doing it level 3 was purchased by centurylink and is part of some subdivision they now call lumen.

There is a tiny chance the game company buys services directly from level3/lumen for that data center but good luck finding someone at a game company customer service dept that has any clue how their data centers are really hooked to the internet.

 

[silver085]

How about Cloudflare Warp?

Cloudflare Warp had the best ms values where checking various sites and server games through CMD, probably because it was still using servers in my country or something like that? Because when checked with "what is my ip" sites, those were still showing my country, just different IP and city, meanwhile other VPN's were just switching it to different countries. What's weird that with Warp, my in game ping was supposedly the same like without any VPN, yet other people were still teleporting, but only slightly and it was kind of consistent more like some frame skipping than actual lag where stuff teleports in random intervals and random distance. If I remember correctly I also had a normal control over my character like my ping wasn't high. Alas I logged out and ran WinMTR with Warp enabled and got ping spikes aswell after some time, though I noticed the route was quicker, If I remember it was 6-7 hops instead of 11.

 

[lantis3]

Cloudflare & Proton both use Wireguard protocol (usually a lot faster than OpenVPN), whether it will improve the situation definitely depends on their routes and their server locations, however.

 

[silver085]

Your problem is you can only fix what you have control over.

This is mostly your pc, the router and to some small extent the connection you pay your ISP for between your house and the ISP.

The data from both test cases uses the exact same devices so it is not likely there is a issue in these devices. What your traces show is there is some issue far away from your house likely in level 3 network. Do you really think you can just call up level 3 and they will change things for some gamer. Even if you were thinking of doing it level 3 was purchased by centurylink and is part of some subdivision they now call lumen.

There is a tiny chance the game company buys services directly from level3/lumen for that data center but good luck finding someone at a game company customer service dept that has any clue how their data centers are really hooked to the internet.

Hey bill, not sure if you remember, but you were helping me back here last year :
https://forums.tomshardware.com/thr...ed-ethernet-connection.3808293/#post-23014153

Short story long, changing router fixed that issue, though now again I have that CGA4236TCH1 and that every 4h disconnect came back with it ... though I could live with it, atleast compared to what's happening right now.

I found it important to mention, because that last year, my ISP was constantly asking for WinMTR reports while trying to fix those disconnects issue. I still have those, some to that very MMO server aswell, so we can compare how it looked before when I didn't have those ping issues. The adress is slightly different, nonetheless it's the same game, they just use few different IP's with last 3 digits changed that I think are changing daily. I get those from task manager -> resources tab -> network and then from game's exe. process.

So this one is from around May last year. There were still some spikes doubling the average, but there were no packet losses which means those were some single random packets that spiked every now and then right?

Basically this problem started to occur right after new year if I remember correctly. So does this mean something has happened since then to ae5-462.bar1.budapest1.level3.net which now causes this issues?

But thing is... it's not only limited to that 1 game that routes through it. It happens on every game, all it takes it just for the server to be located in foreign country. I checked it on 3 games, Elder Scrolls Online(MMO), Left 4 Dead 2, Counter Strike 2. For the latter 2 games, I just took some random German's IP servers from https://www.gametracker.com/servers/ and Polish ones as well to compare it at the same time. In both games the result is always the same. The moment it starts to spike and drop packets on German servers, Polish one remain stable, just like you can see in my pastebin links.

Now what I realized is since I was told about WinMTR, I don't think I ever checked any other server than that MMO's one. I'll use it tommorow on those other 2 games just out of curiosity to see how the route on those look like and where it starts to drop packets. Nonetheless it happens on every game anyways, so it won't probably be useful.

The last important thing to mention is that this problem disappears during the night.

Here's the log from today that I started I think around 11 AM in Europe and just stopped it before 2 AM.

The 1% on the first hop sometimes appear on my PC, but it's unrelated as I don't have it on my laptop when I unplug ethernet cable from PC and plug into laptop for the time being while doing tests. The first 3 tests from my 1st post are from this laptop. Those 1 or even actually 2% as you can see above in the previous screenshot that shows last year were there already, back when this ping spikes issue wasn't happening so it's unrelated.

 

[silver085]

So I checked this server today with WinMTR https://www.gametracker.com/server_info/s3.core-ss.org:27015/

but it's same story as usual in just every game - random ping spikes and packet drops on foreign addresses

 

[lantis3]

Internet traffic is just like car traffic. Traffic condition can be very good around your neighborhood, but once you get out of the area, the condition can be very bad, can you fix it? No. The only "possible" solution is to change route. But which route? You can only test yourself one by one, via VPN, as mentioned above.

My test result from US west coast for the server you provided

 

[silver085]

Internet traffic is just like car traffic. Traffic condition can be very good around your neighborhood, but once you get out of the area, the condition can be very bad, can you fix it? No. The only "possible" solution is to change route. But which route? You can only test yourself one by one, via VPN, as mentioned above.

My test result from US west coast for the server you provided

But how come I was able to play those very same games for like 15 years without any issues? I was able to play any online game without any ping spikes or packet losses. Now it's like I'm condemned only to servers located in my country. If the server is located elsewhere, I'm just risking those issues. Speaking of these 15 years, I mean all these Counter Strike versions that came out over the years for example, or both Left 4 Dead and Left 4 Dead 2 games. I played those since I was 10 and now I am 27. That one server that I just showed report of from today, which you also did, I used to play on it few years ago too without any problems.

I do get your explonation about Internet traffic, but thing is - I didn't change my living conditions which could explain it. I didn't move to a new country, new house, didn't get new ISP and didn't set up new gear. If I did and then noticed these problems from the day 1, it could be explained that I just simply chose wrong ISP for my place, like that ISP has some routing problems and have to try different one. Yet this is same ISP for like 15 years I think? I play that one MMO which I was posting logs from since June 2018 and never had problems with it. This problem started to occur around new year I think and what's weird, why does it affect every foreign server from every game?
You're also saying that the only solution is changing the route, but I also posted log from the last year from that MMO server and the route is pretty much the same back when this problem didn't exist

Today's routing seems to have 1 more extra hop just before final address. To me this seems more like some kind of network overload where it has problems with far away destinantions. Just realized I didn't mention this in the beginning - those spikes are always happening simultaneously when it comes to foreign servers.

For example I was playing that MMO and had 2 cmd pings ran in the background. The moment I start to lag in game, I alt tab to see MMO cmd ping and indeed several 150ms lines with some timed out requests in between, then I quickly check german's Left 4 Dead 2 server cmd ping and it looks just the same.

I had those simultaneous pings running a lot of times, I will link both german Counter Strike and Left 4 Dead 2 servers for example :

L4D2: https://pastebin.com/B75TvMSp
Counter Strike 2: https://pastebin.com/MhZSyKr3
In the latter for some reason ping wasn't spiking, yet the packets were being dropped.

The fact this issue doesn't appear at all during night hours is also weird. Is there some kind of software that could monitor network activity/traffic ?

 

[bill001g]

In the end your only hope is some kind of vpn service but it is all trial and error some of the vpn providers might follow the same route. You never know maybe it would be more stable to open a vpn to a data center in say france and then go back to germany. The over all latency might be bit higher but it is much better than large spikes or packet loss.

Do you know anyone in germany. What happens if you run ping to their house... they likely have to allow ping on their router.
This would mean you might get lucky and there is a vpn data center in germany you could get to that could then connect to the gaming company data center with no issues.

Time of day issue are almost always a overloaded connection. People use more and more bandwidth to say watch 4k netflix. The various isp in the path might not have upgraded some of their connections fast enough. The costs to have more fiber run or to say go from 10gbit to 40gbit is a lot.

It could also be something like someone dug up and broke a fiber bundle and they switched over to a backup connection that is not as large.

In the end there is nothing you can do change or affect the path your data takes. You give your data to the ISP router on the far end of the wire going to your house and they handle the delivery. Kinda like if you mail a letter, you give it to your post man but after that it will go though lots of trucks and offices and maybe even other countries. You have no way control which roads the trucks take.

 

[lantis3]

Peering ISP/partner can change, owner can change, management level can change, tech staff can change, cost cutting measures can be happening, equipment can be dying and become unstable, line quality can deteriorate somewhere, who knows?

You always share the same pipeline with a lot of users, anyone can saturate/disrupt the pipeline at any time/any point. Internet traffic is never guaranteed unless you have a direct line to the server.

---

https://www.google.com/search?q=free+germany+vpn

 

[silver085]

Well, just for the record. It was fine for about a week and not just during the night, but during day as well, it just stopped completely. I have an old report from March 23 where you can see it's fine :

Latency was good and it was just a single packet being lost in few places. The problem returned yesterday, and I did a test today just a few hours ago.

Compared with the one above, this does look like some serious network overload :

So this is my ISP's doing after all?

 

[Ralston18]

"So this is my ISP's doing after all?"

What specifically are you looking at to draw that conclusion?

 

[silver085]

"So this is my ISP's doing after all?"

What specifically are you looking at to draw that conclusion?

The fact that packet losses and high latency spikes are also present in the first half of the hops below level 3 this time.

 

[Ralston18]

Likely outside of your ISP.

Run ping, pathping, and tracert via the Command Prompt on your PC.

Target Google at 8.8.8.8 or 8.8.4.4

Post the results accordingly via Copy and Paste.

 

[silver085]

pathping

Spoiler

Śledzenie trasy do dns.google [8.8.8.8]
z maksymalną liczbą 30 przeskoków:
  0  Sebastian [192.168.0.6]
  1  192.168.0.1
  2  10.236.0.1
  3  172.17.154.1
  4  172.17.64.10
  5  172.17.64.10
  6  142.250.163.248
  7  192.178.97.13
  8  108.170.234.101
  9  dns.google [8.8.8.8]

Wyliczanie statystyk dla 225 sekund...
            Źródło   Ten węzeł/Łącze
Przeskok  RTT    Zgubione/wysłane = Pct  Zgubione/wysłane = adres Pct
  0                                           Sebastian [192.168.0.6]
                                0/ 100 =  0%   |
  1    0ms     0/ 100 =  0%     0/ 100 =  0%  192.168.0.1
                                0/ 100 =  0%   |
  2    9ms     0/ 100 =  0%     0/ 100 =  0%  10.236.0.1
                                0/ 100 =  0%   |
  3  ---     100/ 100 =100%   100/ 100 =100%  172.17.154.1
                                0/ 100 =  0%   |
  4  ---     100/ 100 =100%   100/ 100 =100%  172.17.64.10
                                0/ 100 =  0%   |
  5  ---     100/ 100 =100%   100/ 100 =100%  172.17.64.10
                                0/ 100 =  0%   |
  6   19ms     0/ 100 =  0%     0/ 100 =  0%  142.250.163.248
                                0/ 100 =  0%   |
  7   17ms     0/ 100 =  0%     0/ 100 =  0%  192.178.97.13
                                0/ 100 =  0%   |
  8  ---     100/ 100 =100%   100/ 100 =100%  108.170.234.101
                                0/ 100 =  0%   |
  9   17ms     0/ 100 =  0%     0/ 100 =  0%  dns.google [8.8.8.8]

Śledzenie zakończone.

tracert

Spoiler

Śledzenie trasy do dns.google [8.8.8.8]
z maksymalną liczbą 30 przeskoków:

  1    <1 ms    <1 ms    <1 ms  192.168.0.1
  2    14 ms    27 ms     7 ms  10.236.0.1
  3     8 ms     9 ms     9 ms  172.17.154.1
  4    18 ms    19 ms    17 ms  172.17.64.10
  5    17 ms    17 ms    16 ms  172.17.64.10
  6    17 ms    17 ms    17 ms  142.250.163.248
  7    17 ms    17 ms    17 ms  192.178.97.13
  8    20 ms    19 ms    15 ms  108.170.234.101
  9    16 ms    17 ms    18 ms  dns.google [8.8.8.8]

Śledzenie zakończone.

ping

Spoiler

Badanie 8.8.8.8 z 32 bajtami danych:
Odpowiedź z 8.8.8.8: bajtów=32 czas=18ms TTL=118
Odpowiedź z 8.8.8.8: bajtów=32 czas=17ms TTL=118
Odpowiedź z 8.8.8.8: bajtów=32 czas=19ms TTL=118
Odpowiedź z 8.8.8.8: bajtów=32 czas=16ms TTL=118

Statystyka badania ping dla 8.8.8.8:
    Pakiety: Wysłane = 4, Odebrane = 4, Utracone = 0
             (0% straty),
Szacunkowy czas błądzenia pakietów w millisekundach:
    Minimum = 16 ms, Maksimum = 19 ms, Czas średni = 17 ms

Correct me if I'm wrong, but these would probably be useful when done during the actual latency spikes moments? As of now, all values are acceptable here, because I am not experiencing this issue at this moment.

 

[Ralston18]

Is there anything else connected between your incoming ISP connection and the router?

Do you have a modem and a router as separate devices or as a combined (modem/router) device?

Make and model information?

The router is using a commonly used IP address for routers (Gateway).

I am not sure about the device 10.236.0.1.

= = = =

If the times of latency spikes are long enough for you to become aware of and subsequently run the commands then try running the commands.

Also (late edit) when you mentioned a single packet being lost are you referring to the % column?

 

[bill001g]

When it is random you need a one of these fancy ping check tools runs constantly but most only run for a small amount of time.
You are in effect going to have to make your own. Open a number of cmd windows and leave constant ping run to some of the IP. You could do all of them in the path but better if you select ones you suspect based on your first ping test that showed issues.

The 3 minimum should be your router, the ISP first router and the final test server. You can select more to try to narrow it down.

You are correct if we just look at the first test you posted. You see packet loss starting in hop 2 but it is only a single packet but starting in hop 3 and beyond you are loosing about 20 packets. These are not the same exact packets since each hop is tested individually but if hop 3 is dropping 1% of the packets going to hop 4 it will also drop 1% of the packets going to hop 5.

1% is actually a huge amount when it comes to online games. What tends to happen is it will lose a couple in row which is worse than losing them on a more even basis even though the average rate is still the same.

So this is likely your ISP. It is not the connection between your house and the router in hop 2. It could be the hop 2 router , the connection between hop 2 and hop 3 or the hop 3 router. Problem is only the ISP can see or test this. You can not even be 100% sure the data going from router 2 to router 3 follows the same path as the return data going back from router 3 to router 2.
The ISP may have redundant connections which could also explain why your issue is random.

In any case you want to try to collect more test data that shows this. It needs to very consistently look like your first test. When the problem is random sometime the tools like this lie if the problem comes and goes during the testing.


Then the hard part is you are going to have to understand this well enough to explain it to the level 1 tech you talk to . The level 1 guys really only understand the connection to people houses and trouble shooting dumb user setup issues. They likely have no access to any of this equipment so your goal is to get to a person that does.
Packet loss like this is actually a good thing, kinda of. If you have very high latency spikes but no loss then it is more likely load related. Packet loss although it can be load releated tends to be some kind of defective equipment. That the ISP can easily fixed those, a issue like they are exceeding the bandwidth on a port on their router is going to be much harder for them to fix than just replacing some defective device or cable.

 

[silver085]

Is there anything else connected between your incoming ISP connection and the router?

Not that I am aware of. It's just my device that I got from my ISP that is connected directly to my PC with ethernet cable.

Do you have a modem and a router as separate devices or as a combined (modem/router) device?

It's a combined Modem & Router device from what I know.

Make and model information?

Technicolor CGA4236TCH1

If the times of latency spikes are long enough for you to become aware of and subsequently run the commands then try running the commands.

Also (late edit) when you mentioned a single packet being lost are you referring to the % column?

No, I was just referring to Sent vs Received packet nubmers. The % column is pretty much always showing 1%, because I always run those tests for a long time. Let's say 10 packet being lost in a total 3/5/10k sent will just be shown as only 1% right? While talking about that first WinMTR report from March 23, where it was one of those days the connection was stable, you can still see 1% lost, but if you look at the numbers, it's just 1 random packet that was being lost on a few hops. Nothing to worry about, as latency values were stable too. The test itself was few hours long too, as it can be seen by 9k sent packets.

Then if you look at the report from today, it still shows only 1%, however looking at Sent vs Recv numbers, in a short test (less than 4k packets) it already shows over 20 lost packets in all hops, as well as high latency spikes everywhere. This is how it looks like while doing ping command :

Spoiler

No one in this thread mentioned pathping and tracert commands before and those spikes usually last for around 10-15 seconds, in which case I suppose I can prepare and type in those commands in CMD, play the game as usual and the moment I'll be experiencing this issues in game, I will quickly alt tab and hit enter in CMD in hope of capturing something.

 

[silver085]

When it is random you need a one of these fancy ping check tools runs constantly but most only run for a small amount of time.
You are in effect going to have to make your own. Open a number of cmd windows and leave constant ping run to some of the IP. You could do all of them in the path but better if you select ones you suspect based on your first ping test that showed issues.

The 3 minimum should be your router, the ISP first router and the final test server. You can select more to try to narrow it down.

You are correct if we just look at the first test you posted. You see packet loss starting in hop 2 but it is only a single packet but starting in hop 3 and beyond you are loosing about 20 packets. These are not the same exact packets since each hop is tested individually but if hop 3 is dropping 1% of the packets going to hop 4 it will also drop 1% of the packets going to hop 5.

1% is actually a huge amount when it comes to online games. What tends to happen is it will lose a couple in row which is worse than losing them on a more even basis even though the average rate is still the same.

So this is likely your ISP. It is not the connection between your house and the router in hop 2. It could be the hop 2 router , the connection between hop 2 and hop 3 or the hop 3 router. Problem is only the ISP can see or test this. You can not even be 100% sure the data going from router 2 to router 3 follows the same path as the return data going back from router 3 to router 2.
The ISP may have redundant connections which could also explain why your issue is random.

In any case you want to try to collect more test data that shows this. It needs to very consistently look like your first test. When the problem is random sometime the tools like this lie if the problem comes and goes during the testing.


Then the hard part is you are going to have to understand this well enough to explain it to the level 1 tech you talk to . The level 1 guys really only understand the connection to people houses and trouble shooting dumb user setup issues. They likely have no access to any of this equipment so your goal is to get to a person that does.
Packet loss like this is actually a good thing, kinda of. If you have very high latency spikes but no loss then it is more likely load related. Packet loss although it can be load releated tends to be some kind of defective equipment. That the ISP can easily fixed those, a issue like they are exceeding the bandwidth on a port on their router is going to be much harder for them to fix than just replacing some defective device or cable.

Right, I will do individual pings to some of addresses too. Probably will do few WinMTR test too for a couple of days in order to be sure, if this is some kind of yet another new problem my ISP is serving me since yesterday where problems are occuring on all hops(vs to what we had in the beginning of this thread with issues being in the adrdesses outside of my country) or maybe it was just some random 1-2 days of ISP having some global issues on their side. I just wanted to know if I should already sent this report from today to my ISP, which is indicating problems on every hop now, but guess I will try to collect some more informations and narrow it down further first, assuming this new problem with all hops is to stay.

 

[bill001g]

This is a much better issue to have than it really only being to international servers. In this case all the equipment and the confirtuation is in control of a single ISP and you are there actual customer who is paying them money. If it was a different ISP they would not care about you. International connections also tend to be the place that a ISP might not purchase enough bandwidth from other partner ISP. Technical issues you at least have a chance to get it fixed business things not so much.

 

[silver085]

For the last few days it's been alright, and in the last 2-3 it's been happening again. This time again it seems to be this "global" issue because everything was slowing down. I was watching a livestream when it suddenly started buffering, other web sites were very slow to load and once I pinged a news web site of my country - www.wp.pl of which I always have 25ms, it was around 90ms during that time.

Today while playing I started to lag again, alt tabbed quickly and pinged that web site from my country to check first if it is that global issue again. If the latency would be stable, then it would mean that this foreign servers issue came back instead.


It was global problem again as this is ping to a web site from my country www.wp.pl which I usually have 25ms to.

It went like this for around 30 seconds I think, so I managed to do both tracert and path ping to 8.8.8.8 in the meanwhile