Frequent Visitor
•
8 Messages
Network issues to certain servers (German servers)
I saw another post recently about issues with connecting to Germany. My average ping to Germany based servers was always 120-130 which is perfectly fine for me. (I need to connect to a specific AWS (amazon server) based in EU, Germany for work). I tried everything on my end, multiple PC's, refresh signals, restarting modem (ARRIS SB6190), and restarting router. I tried hardwiring directly to my modem to ensure it wasnt my router and recieved the same issue. Finally, I tried one last thing. I ran the same tests but with a VPN. The moment I used the VPN, everything was fixed and there were no more packets being lost along with normalized ping below 130ms. This has been happening on and off for weeks now. Any helpful suggestions or fixes? I really would hate having to switch ISP as this is desperately needed. If you need any other information, I will check this post ASAP with any info needed.
https://i.imgur.com/cb0MFW2.png (list of IP's with issues)
https://i.imgur.com/uGr7iEF.png(CMD pinging IPs)
https://i.imgur.com/PODukw2.png (Ping of all AWS servers to show German servers are impossible to connect to, compared to other servers.)
Tracing route to ec2-3-120-0-0.eu-central-1.compute.amazonaws.com [3.120.0.0]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms 192.168.1.1
2 10 ms 11 ms 9 ms 96.120.97.181
3 11 ms 18 ms 9 ms 68.86.163.153
4 10 ms 10 ms 8 ms 96.110.4.201
5 11 ms 10 ms 9 ms 96.110.4.229
6 11 ms 11 ms 10 ms lag-8.ear1.Miami1.Level3.net [4.68.74.237]
7 * * * Request timed out.
8 122 ms 121 ms 121 ms 212.73.243.150
9 * * * Request timed out.
10 * * * Request timed out.
11 * * * Request timed out.
12 * * * Request timed out.
13 * * * Request timed out.
14 * * * Request timed out.
15 * * * Request timed out.
16 * * * Request timed out.
17 * * * Request timed out.
18 177 ms 161 ms 162 ms 54.239.5.164
19 * * * Request timed out.
20 * * * Request timed out.
21 * * * Request timed out.
22 171 ms 154 ms 159 ms ec2-3-120-0-0.eu-central-1.compute.amazonaws.com [3.120.0.0]
That's just one tracert example, usually they are MUCH worse and its for every IP listed in the image links I posted, sometimes reaching 170, 200, 500+ pings. During each of these times when it acts up most, nothing else is wrong with the connection everywhere else. And its only with comcast that this is happening.
As I stated earlier, using any VPN always fixes the issue, meaning the issue is NOT on Amazon Server's end. Its an issue with comcast. I hope this is urgently fixed, as this is an urgent matter.
(I saw another customer post similar issues as me https://forums.xfinity.com/t5/Your-Home-Network/Accessing-some-sites-in-Germany-is-slow/td-p/3287229 whether its related or not, it clearly is an issue affecting work. If there's an employee who can look into this that would be great, as the call reps keep insisting refresh signals and such, even with all my independant testing/results/proofs.)
Edit: A 3rd party network technician told me only a higher up would be able to fix this. Since the issue is fixed via VPN, its a specific routing issue or an internal issue with ISP itself. In other words, nothing from my end will fix it and only comcast can fix this.
Also here is a picture of realtime chart to show what happens on hops PLEASE view this specifically as well. https://i.imgur.com/HESGDIo.png
jsdoug22
New Poster
•
7 Messages
5 years ago
0
0
falconeye
Frequent Visitor
•
8 Messages
5 years ago
I went ahead and used a program to check latency for hetzner as well. You're right. Its not as bad as the AWS EC2 Frankfurt server but its substansial enough that I would cry having that too. Not sure what's been going on. But hopefully an employee comes and see's all this. Each arent even recieving packets.
0
0
EG
Expert
•
111.5K Messages
5 years ago
FWIW; that trace shows that the packets are beyond Comcast webspace and are now into the backbone transit webspace of the level3 provider. It is they who should be contacted concerning this. The Comcast local and regional routers all have low latency.
0
falconeye
Frequent Visitor
•
8 Messages
5 years ago
My tech friend said ISP paths the route taking into account budget/performance, and that they could shift the entire route with miniscule changes, just as I was able to do myself as I ended up using something else (similar to a vpn), to create direct pathing to specififc nodes. Tier 3 is already slow networking over long distance, but it is indeed most budget friendly as there is wide variety of pathing choices. Just because comcast's end is fine, doesnt mean that they still cant make a tweak to fix this though via routing. As just a consumer though, if the company has the power to fix it with a minor tweak by taking an altered route, I think they should do it. After all, we are using their service.
On top of that, the routing should be based on your starting point until it leads to the end server. Others are having the same issue as me who have much different starting points than me. Coincidence that comcast is pathing the same route even though we are different geo locations? Pathing ideally takes the most effective route to nodes etc, but it seems more like they take the most cost effective route compared to performance.
0
0
EG
Expert
•
111.5K Messages
5 years ago
It's beyond the scope of this forum to discuss BGP routing metrics. Research into the subject is suggested.
Route selection has no metric for geographical distance or latency -- routes are selected purely based on the least number of carrier transitions. With BGP routing, the routing is determined pretty much by the remote reciever, not the client / sender.
Ex: Comcast (TX) -> TWTC (CA) -> CyrusOne (TX)
is a shorter route than
Comcast (TX) -> Level3 (TX) -> TWTC (TX) -> CyrusOne (TX)
That's just the way the Internet works.
Comcast pretty much only has one external route: 0.0.0.0/0
They pass everything to their upstream bandwidth provider for a given market (Level(3) and ATT are a couple of examples, and that provider works out how to forward the traffic based on a number of factors. A few years ago it was just standard BGP metrics that influenced the decision, but today nearly all backbone providers are using technology like intelligent route reflectors to engineer their traffic flows for better performance and to avoid point congestion. This can result in paths being used that are not latency-equivalent, but provide much better service in terms of capacity and packet loss.
The only thing Comcast can influence is which carrier network onto which a packet will be forwarded. After that they are unable to control any aspect of which path will be chosen. Once it's on the upstream carrier's network all routing decisions are firmly in that carrier's hands.
Routing is even way more complicated than this.
0
0
BEAZED1
New Poster
•
2 Messages
5 years ago
Hi,
I had a routing change made earlier today can you check if this is still an issue?
@falconeye
Thanks
0
0
jsdoug22
New Poster
•
7 Messages
5 years ago
The issue had resolved itself a few days ago and now it is back, someone please address this
0
0