Dear community - first of all, super useful forum, really like the information shared!
Let me start with the problem statement: I think the latency for the type of connection I having (Fiber) is way too high. I have traced it back to the link between my home and the next hop (presumably Point of Presence, literally 1.5 km away from my home). I am seeing about 5 ms to that first hop, and hence everything else has increased latency. What are your experiences on fiber?
A bit more background: I live in Leidsche Rijn and number of years ago ReggeFiber came by installing fiber on my street and respectively I was the first one to sign up. I got the internet subscription back then from XS4ALL, and that was pretty much until KPN took over. Then I switched to T-Mobile Thuis over the same fiber. Lots of issues with the standard modem and media convertor etc, so at some point of time got sick of it and threw everything away, moving completely to UniFi solutions - main router being UDM-Pro.
So the setup right now is:
Fiber from ReggeFiber directly into UDM-Pro, from there fiber to my switch etc.
Now, as you know ReggeFiber got acquired by KPN at some point of time, so something strange happened… The latency to my next hop from sub 1 ms went up to almost 5 ms…
Called the service desk number of times and heard all kind of responses, like:
- What is latency?
- If you can reach ”internet” what is the problem”
- etc etc - you get the picture
At some point of time I even managed to get a person to actually come into my house and check everything. That as well didn’t go so well, as the guy said: I am only hardware tech, sorry - if any hardware is broken I can fix/replace it. Fiber connection to the PoP is not his responsibility… Needless to say, he insisted to test everything with T-Mobile equipment so I had to resurrect that from the dead, but surprise, surprise - the latency was pretty much the same as with my UDM-Pro. He said - not my problem and went away...
I’ve been with T-Mobile Thuis for number of years already, I’ve no contract obligations etc.
My plan was always 500/500. And to be honest I’d give half of this for a lower latency!
I have a direct comparison with a friend on KPN fiber, and he gets consistently less than 1 ms to the next hop…
Here are some examples (Executed on UDM-Pro, where fiber terminates):
# traceroute 1.1.1.1
traceroute to 1.1.1.1 (1.1.1.1), 30 hops max, 46 byte packets
1 xx-xx-xx-62.ftth.glasoperator.nl (62.xx.xx.xx) 4.426 ms 4.450 ms 4.394 ms
2 one.one.one.one (1.1.1.1) 8.382 ms 8.002 ms 8.092 ms
traceroute to 8.8.8.8 (8.8.8.8), 30 hops max, 46 byte packets
1 xx-xx-xx-62.ftth.glasoperator.nl (62.xx.xx.xx) 4.521 ms 4.499 ms 4.577 ms
2 dns.google (8.8.8.8) 8.336 ms 7.578 ms 8.052 ms
# traceroute ams-ix.net
traceroute to ams-ix.net (185.55.136.60), 30 hops max, 46 byte packets
1 xx-xx-xx-62.ftth.glasoperator.nl (62.xx.xx.xx) 4.467 ms 4.471 ms 4.314 ms
2 lb.ams-ix.net (185.55.136.60) 9.411 ms 9.728 ms 9.522 ms
# ping ams-ix.net
PING ams-ix.net (185.55.136.60): 56 data bytes
64 bytes from 185.55.136.60: seq=0 ttl=60 time=11.077 ms
64 bytes from 185.55.136.60: seq=1 ttl=60 time=10.850 ms
64 bytes from 185.55.136.60: seq=2 ttl=60 time=10.938 ms
64 bytes from 185.55.136.60: seq=3 ttl=60 time=10.876 ms
So, back to my question: what are your experiences with your latency to the next hop (PoP)? I am having a hard time accepting 5 times increase...