|
|
I see 73 down, 21 up, and 13ms minimum latency (more like 17ms to most destinations). I'm not complaining, but I'm also not sure why the downstream isn't hovering right around the cap as the upstream is or why the latency to the first hop* isn't very low single digits instead of 12ms. What do you get on your connection?
(I'm pretty confident my computer and local network equipment are OK, not looking for troubleshooting steps as such. Just want to determine whether this is typical for Infinity 2 FTTP)
* Actually the very first hop doesn't respond to icmp in my case, so really the second hop.
Edited by deleted (Mon 23-Oct-17 07:30:48)
|
|
|
Hmm, FWIW, a tad over the ordered upload speed is what I have observed on FTTP, with the download usually being a couple of meg less .......
not noticed the ping being into double figures, but then that may depend on where in the country you are possibly ?
Which router are you using ?
|
|
|
Okay, it's probably in the right ballpark then. Although I would have expected something closer to 77 than 73 given that info.
Realistically the speed is fine at the moment. The reason I wonder is that the line is supposed to be able to do 330+Mbit, so it seems strange that it can't do a locked 80 or very close to it.
Currently just using the ISP's router (BT Smart Hub). I suppose I could try bypassing any router and see what a direct PPPoE connection gives...
I'm in semi-rural Scotland but I should clarify that most of the latency is coming from the first couple hops into the network, not from the hops between BT and the final destination. I would have expected that portion of the latency to be low and not area dependant.
Edited by deleted (Mon 23-Oct-17 08:59:24)
|
|
Register (or login) on our website and you will not see this ad.
|
|
|
Hopefully I'll be able to try it in a few weeks and let you know. The FTTP service has been ordered at 80/20 for now and I'm waiting for an install date.
|
|
|
The reason I wonder is that the line is supposed to be able to do 330+Mbit, so it seems strange that it can't do a locked 80 or very close to it.
That is the highest profile currently offered as a product on your service, the technology could do more.
I daresay someone will be along shortly to explain about why one doesn't get the full kahuna downstream, packet headers and the like ....
Paul is on a 330 Openreach FTTP product, there's a speediest posted in the linked post.
|
|
|
Paul is on a 330 Openreach FTTP product, there's a speediest posted in the linked post.
I normally get a little more than what was displayed on that image.
But the most I have got is 320 Mbit down and 35 Mbit up and the latency has been as low as 3ms but averages about 9ms.
BT's Infinity 4 is listed as 300 Mbit down and most of the time I get more but there are times especially between 3pm and 2am week nights when it drops below 200 Mbit, but its still better than the 4-5 Mbit I was getting on ADSL2+ isn't it 
The latency to this forum is shown below:
ping thinkbroadband.com
Pinging thinkbroadband.com [80.249.99.72] with 32 bytes of data:
Reply from 80.249.99.72: bytes=32 time=4ms TTL=53
Reply from 80.249.99.72: bytes=32 time=5ms TTL=53
Reply from 80.249.99.72: bytes=32 time=4ms TTL=53
Reply from 80.249.99.72: bytes=32 time=3ms TTL=53
Ping statistics for 80.249.99.72:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 3ms, Maximum = 5ms, Average = 4ms
A trace route to this site:
tracert thinkbroadband.com
Tracing route to thinkbroadband.com [80.249.99.72]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms *ROUTERNAME* [192.168.2.253]
2 3 ms 2 ms 2 ms 217.32.146.68
3 3 ms 2 ms 2 ms 217.32.146.110
4 5 ms 3 ms 3 ms 217.32.147.194
5 4 ms 3 ms 4 ms 213.120.178.67
6 3 ms 4 ms 3 ms 217.41.168.107
7 4 ms 4 ms 4 ms 109.159.249.106
8 4 ms 4 ms 4 ms core4-te0-10-0-26.faraday.ukcore.bt.net [109.159.255.197]
9 5 ms 4 ms 4 ms 213.121.193.199
10 5 ms 4 ms 4 ms linx-gw1.thn.ncuk.net [195.66.224.240]
11 6 ms 4 ms 4 ms te2-1-9.star10g.bdr-rt3.thdo.ncuk.net [80.249.97.17]
12 6 ms 6 ms 8 ms po4-31.core-rs2.thdo.ncuk.net [80.249.97.85]
13 4 ms 4 ms 4 ms ip99-72.thdo.ncuk.net [80.249.99.72]
Trace complete.
The bbc site is eactly 8ms, google.com is 4ms to 5ms, our server located in France is about 10ms to 12ms.
The World of Warcraft Europe Servers are about 13ms, Diablo III Europe Servers settles down to about 18ms after a couple of mins.
So all in all its not too bad
A neighbour across the road from me is also on the 80/20 package from BT over FTTP and he gets the same latency as me and he gets about 74 Mbit down, 20Mbit up and that was over Wi-Fi (10m away from router), so I am not sure why the OP is getting such a higher latency.
Paul
|
|
|
I am not sure why the OP is getting such a higher latency.
As the OP is in Scotland he's never going to get pings in single digits, assuming his ISP's servers are in England. On my BT 21CN FTTP line I never get less than 19ms which is about right considering I'm in the Highlands and my ISP's servers are in London. On FTTC (TalkTalk) I never got less than 22ms. You can't beat the laws of physics!
|
|
|
Thanks for the stats, PaulKirby.
baby_frogmella: The thing is, almost all of it is in the first hop (or first two). There are many hops after that where the packet works its way through a long network, but those add very little delay.
I don't know where my PPPoE link ends and regular traceable packet hopping commences, but it seems unlikely to be in London.
On my BT 21CN FTTP line I never get less than 19ms
How much of that is from the first hop after your router?
|
|
|
Here's mine (reformatted slightly, macOS' traceroute is less readable than Windows'):
traceroute to thinkbroadband.com (80.249.99.72), 64 hops max, 72 byte packets
1 192.168.1.254 (192.168.1.254) 1.123 ms 0.849 ms 0.744 ms
2 * * *
3 31.55.185.197 (31.55.185.197) 11.984 ms 11.871 ms 11.734 ms
4 31.55.185.220 (31.55.185.220) 11.766 ms 12.026 ms 12.010 ms
5 195.99.127.114 (195.99.127.114) 12.188 ms 11.982 ms 12.162 ms
6 195.99.127.87 (195.99.127.87) 12.487 ms 12.367 ms 12.061 ms
7 linx-gw1.thn.ncuk.net (195.66.224.240) 12.391 ms 12.372 ms 12.306 ms
8 te2-1-9.star10g.bdr-rt3.thdo.ncuk.net (80.249.97.17) 12.134 ms 12.126 ms 12.504 ms
9 po4-31.core-rs2.thdo.ncuk.net (80.249.97.85) 12.777 ms 12.519 ms 13.116 ms
10 ip99-72.thdo.ncuk.net (80.249.99.72) 12.201 ms 12.263 ms 12.165 ms
As you can see, the all-important first hop doesn't respond to pings. If it were 3ms to that and 9ms from there to the second hop, I would understand that and assume it was distance related. As things are, though, I suspect it's more like 9-10ms to the first hop which is much higher than GPON should provide.
|
|
|
This is the traceroute to TBB from my FTTP 330/30 connection:
C:\>tracert thinkbroadband.com
Tracing route to thinkbroadband.com [80.249.99.72]
over a maximum of 30 hops:
1 1 ms 1 ms 1 ms 192.168.1.1
2 24 ms 21 ms 22 ms 77-89-165-7.fluidata.co.uk [77.89.165.7]
3 44 ms 21 ms 21 ms 77-89-165-29.fluidata.co.uk [77.89.165.29]
4 24 ms 22 ms 22 ms linx-gw1.thn.ncuk.net [195.66.224.240]
5 24 ms 21 ms 22 ms te2-1-9.star10g.bdr-rt3.thdo.ncuk.net [80.249.97.17]
6 25 ms 24 ms 22 ms po4-31.core-rs2.thdo.ncuk.net [80.249.97.85]
7 25 ms 21 ms 21 ms ip99-72.thdo.ncuk.net [80.249.99.72]
Trace complete.
As you can see, i more or less get a constant 22ms from the second hop (first hop is always a local hop to your router). The above is through 802.11ac wifi so wired resultts will be a few ms better.
TBH unless you're having trouble with your connection, I wouldn't worry too much about tracert results. Setting up a TBB BQM will give you a better idea of your latency and as long as its smooth you've got nothing to worry about. But I'll re-iterate: your geographical location plays a big part in your ping times.
Edited by deleted (Tue 24-Oct-17 12:30:36)
|
|
|
|
Also you'll never get the full 80Mbps download speed on the 80/20 FTTP tier, it will hover around the 74-75 Mbps mark due to overheads. Likewise I never get the full 330 Mbps, its around 310 Mbps but never less than 300 Mbps.
|
|
|
|
Traceroute is not a good tool to use when looking for network delay. Ping is better.
Once you've used traceroute to list the intermediate routers you're better off using ping to establish trip times.
Traceroute uses packets will carefully set hops to live settings and waits for the appropriate router to drop the packets and report the drop back to traceroute. Routers often don't prioritise this reporting so traceroute timings can be odd.
In our example a packet to hop 3 took 44ms but one to hop 4 took 24ms. This difference would in part be due to the router in hop taking more time to report the dropped packet.
What pings times to hop 2 do you get?
|
|
|
|
j314
Paul is in London and you can see that hop8 is Faraday Core router near ST Pauls. Very short distance.
I would suggest that your ISPs servers are also in London (taking into account the differences up to the common gateway server at hop 10 on Paul's, 7 on yours and 4 on Baby's) but you haven't said your ISPs name so we can't check that out.
Simplified explanation:-
As has been said this is purely distance related, and the number of (hidden) Network routers that each add a small delay.
Scotland to London at over 1000km adds 0.005 ms delay per km due to lightspeed so 5ms is sheer distance at a minimum.
All the network hops within BT to your ISP handover site are hidden within the encapsulated pipe used to provide the service to your ISP. ie the whole pipe is routed and packets are not switched at intermediate points so no router can respond to pings. That is why Baby's first ping is from Fluiddata in London even though he is in the Highlands.
|
|
|
Where your PPPoE will end depends on your ISP and for many that is actually in London
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
|
Thanks all for the responses. When I say "first hop" I meant the first outside of my home network, which I know isn't strictly correct. I'll use the correct hop numbers from now on. Pinging my router (a BT Smart Hub) is always ~1ms.
bsdnazz:
Good to know about ping vs traceroute, I'll keep that in mind. Hop 2 doesn't respond. Pinging hop 3 (31.55.185.197) gives timings almost the same as traceroute (11-12ms), so I don't think it's a major culprit in this case.
kitkat, MrSaffron:
ISP is BT Retail. If the PPPoE doesn't end until London then I agree that the latency figures are about right (maybe still a bit higher than can be explained using only a speed-of-light delay, but not much of a concern). Is there any way I can confirm that location?
|
|
|
Based on time from our network in docklands to where the first IP hop that is not your network which is 1.2ms I'd say you are emerging in London which is normal.
Remember its speed of light in fibre, and every switch/router even adds a small delay.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
Is there any way I can confirm that location?
What does this give you for 'city'?
http://ip-api.com/
For my connection it gives Poplar, London.
|
|
|
What does this give you for 'city'?
Keswick, a tiny town in Cumbria.
I tried some other geolocation sites while I was at it:
* maxmind's GeoIP2 demo says Wigton, Cumbria
* geoiplookip.net doesn't detect a city but its map also shows Cumbria
* ultratools shows Barnet, London
* geoiptool.com shows my actual location in southern Scotland, down to the correct first half of postcode (!)
Looks like I can't reliably detect the PPPoE exit point from the IP, then. Thanks for the idea though...
|
|
|
Geo IP is almost worthless in the UK, it works for some and fails terribly for others
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
Remember its speed of light in fibre,
A lot less than speed of light.
The speed of light in a vacuum is 299,792,458 meters per second, or 186,282 miles per second. In any other medium, though, it�s generally a lot slower. In normal optical fibers (silica glass), light travels a full 31% slower.
|
|
|
Speed of light is an unknown number unless you name the medium it is travelling in I did.
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
|
|
|
Ah, fair enough. I suspect most (including me) wouldn't have picked up on that.
|
|
|
|
Replying to self to finish this matter off...
First, I've pretty much confirmed that my connection comes out in London by using Ookla Speedtest's manual server selection feature. It autoselects servers near Cumbria based (I assume) on my geoip, but as I force servers closer to London, the ping goes down. Servers in London itself give ~11ms RTT, which seems to be the absolute limit for me.
Second, I fired up a London DigitalOcean VPS and tested latency to various servers in the north of England (no servers in the south of Scotland, but close enough. Used speedtest-cli since it was the easiest way to ping particular locations). Those came out as around 11ms too, so it looks like this time is quite normal.
I'm sure sufficiently determined parties could improve this somewhat by using longer reaches of dedicated fibre/better routing/faster routers, since it isn't at the physical speed-of-light-in-fibre limit yet. But it's reassured me that there doesn't seem to be anything wrong with my connection.
A PPPoE connection that ended earlier than London would allow lower latency to servers north of London, but in practice most of my traffic would likely go through there anyway, so it works out well enough.
Thanks all for the info and advice
|
|
|
If you have a need to emerge at a location closer to you, consider some of the smaller providers who may have their core in Manchester
|
|
The author of the above post is a thinkbroadband staff member. It may not constitute an official statement on behalf of thinkbroadband.
|
Pages in this thread: 
Print Thread
