FTTP pings

faceman123

faceman123

faceman123

Soldato
Joined
9 Jun 2011
Posts
3,639
Hi folks, just wanted to check what sort of pings everyone is getting with FTTP - I'm currently hovering around the 12ms range =\ ( expected it to be a lot lower) is this normal? installation was just yesterday, is there any Teething period with FTTP?
 
Ping to where? I wouldn't expect my latency to be considerably lower with fibre than with a copper connection really.
 
A round trip from Edinburgh to London at the speed of light and in a direct line would take about 4ms, so only losing 8ms to the routing and so on seems fine to me.

Zefan said:
Ping to where? I wouldn't expect my latency to be considerably lower with fibre than with a copper connection really.
Click to expand...
Yarp. I was on FTTC in central London and was pinging about 6ms to bbc.co.uk, etc., and now get 1ms on Hyperoptic. Taking the last mile of copper out of the equation doesn't do much for pings as 99% of the journey was already via fibre.
 
TJM said:
A round trip from Edinburgh to London at the speed of light and in a direct line would take about 4ms, so only losing 8ms to the routing and so on seems fine to me.


Yarp. I was on FTTC in central London and was pinging about 6ms to bbc.co.uk, etc., and now get 1ms on Hyperoptic. Taking the last mile of copper out of the equation doesn't do much for pings as 99% of the journey was already via fibre.
Click to expand...

hi thanks for this - what does 8ms to routing mean? (noob alert) do you mean the router my pc etc or ???
 
Physics determines how quickly the light gets down the cables. I'm not aware of many ISPs that don't send traffic down to London and back - the UK just isn't generally big enough for it to be worth doing any other way.

It might be the case that two customers in Edinburgh on the same ISP will see their traffic take a shorter path to each other, but for traffic going outside the ISP network 99% of the time that peering is happening in docklands somewhere.

For what it's worth, a ping from Loughborough to those two destinations is:

Code: 
gateway # execute ping bbc.co.uk
PING bbc.co.uk (151.101.128.81): 56 data bytes
64 bytes from 151.101.128.81: icmp_seq=0 ttl=61 time=4.6 ms
64 bytes from 151.101.128.81: icmp_seq=1 ttl=61 time=4.5 ms
64 bytes from 151.101.128.81: icmp_seq=2 ttl=61 time=4.5 ms
64 bytes from 151.101.128.81: icmp_seq=3 ttl=61 time=4.5 ms
64 bytes from 151.101.128.81: icmp_seq=4 ttl=61 time=4.5 ms

--- bbc.co.uk ping statistics ---
5 packets transmitted, 5 packets received, 0% packet loss
round-trip min/avg/max = 4.5/4.5/4.6 ms

gateway # execute ping google.com
PING google.com (216.58.206.46): 56 data bytes
64 bytes from 216.58.206.46: icmp_seq=0 ttl=58 time=5.3 ms
64 bytes from 216.58.206.46: icmp_seq=1 ttl=58 time=5.3 ms
64 bytes from 216.58.206.46: icmp_seq=2 ttl=58 time=5.3 ms
64 bytes from 216.58.206.46: icmp_seq=3 ttl=58 time=5.3 ms
64 bytes from 216.58.206.46: icmp_seq=4 ttl=58 time=5.3 ms

--- google.com ping statistics ---
5 packets transmitted, 5 packets received, 0% packet loss
round-trip min/avg/max = 5.3/5.3/5.3 ms


This is on a fibre ethernet service at an on-net exchange. Edinburgh is 3.5x further from London, and your pings are not 3.5x the above. I'd say they are good.
 
Last edited:
You wouldn't notice the difference, 10ms is one hundredth of a second. The only time it might be relevant is if you're pushing a protocol that has no business being used outside a LAN (e.g. pre-3.0 SMB).
 
That was my thought on it BTW, reading it back sounds like I'm getting paranoid! The other thing to consider with ping times is that they can vary throughout the day, but even then it probably isn't noticeable unless there is a big event online that has a DDoS effect on a site (I seem to remember that happening once, can't remember where/when though).
 
ICMP traffic is generally regarded as low priority so many switches and routers will forward other types of traffic before dealing with a ping.

As long as the response times aren't massively high and speed tests check out ok, chances are the response times you see from a ping request won't accurately reflect the response you'd get when transmitting data over a different protocol.
 
Sort of. Ping requests are sometimes processed at a low priority by the *device* at the far end, so you might see ping requests just get lost if the CPU of the device is maxed out, but routers along the path won't be inspecting the traffic flowing through them at several gigabits per second and performing some sort of QoS on them, the aim is to shift packets out another interface and on the way to another router as soon as possible. There's no reason to believe as a rule that the result of an ICMP ping (assuming you get a response) isn't an accurate reflection of the end to end latency.
 
faceman123 said:
Hi folks, just wanted to check what sort of pings everyone is getting with FTTP - I'm currently hovering around the 12ms range =\ ( expected it to be a lot lower) is this normal? installation was just yesterday, is there any Teething period with FTTP?
Click to expand...

Why are people replying to this? 12ms with latency, fttp or not its well you know. 12 flipping milliseconds.
 
If you are super bothered then it might be worth seeing if a provider that peers in LINX Scotland would route broadband traffic to there or if it would still go via their core

https://www.peeringdb.com/ix/745

The actual impact is going to be negligible - maybe you could ping the University of Edinburgh a bit quicker.
 
You must log in or register to reply here.
Back
Top Bottom