Probably...

We don't have any server kit in London.. but do have some routing gear. We'd open a London location but that already has a few.

Matt

Maybe Interleaving has been switched on (for your connection) and that's adding a bit (although I wouldn't expect it to double, or triple the ping delay).

I'm getting 10ms to the Manchester server and 26ms to the MK and Maidenhead servers.
I'm also getting 10ms to Gloucester, Coventry & Birmingham (all non-Xilo/Uno servers), so it looks like BT are routing my connection to the north-east (I'm about the same distance from Coventry and MK, and about the same distance from Maidenhead and Birmingham).

Maybe BT has just changed your routing (entry point onto the www) and it's coming out nowhere near Manchester (what about your pings to other Speedtest.net servers - what if you let it find the closest - what is the ping to that?).

Edited by adebov (Tue 08-May-12 22:01:36)

Hi,

Pingtest reckoned that Maidenhead was the best at 28 ms ... (I'm in York, and the region based shopping sites etc think I should shop in Newcastle - which I suppose is where the Ip appears to be attached to the backbone ???)

Most Speedtest pings come back at 35ms more or less wherever I choose to test. The Manchester one give me the most consistent speed of around 55/13 megs.

The ping though is a bit worrying. Was 10-15 ms from November 2011and then 25ms for the period 25/4 to 5/5 and then has been at 35ms since then.

I changed to 80/20 on 16/4 ... the last time I had the good pings was 13/4 (ie one day after I oprdered the upgrade!)

No coincidence I suspect. Also my IPP has dropped to 57/20 from 66/20 over the last couple of weeks.

On the 40/10 service I was getting 37.5/8 speeds, but now and then had the dreaded slow down until a hub reconnect was done.

On the 80/20, I am seeing 55/13 over my homeplugs - so can't complain - but I have had one instance of a slowdown to 0.5megs until I reconnected the HH3.

Cheers,

Rob

In reply to a post by adebov:

Thanks, so it's in Manchester and whoever posted that it was mislabelled (and is in London) has got their wires crossed (or simply made it up).

Neither actually, it was mislabelled, I reported it on the 20th March based on the following traceroute:

C:\Users\Administrator>tracert 91.199.78.82

Tracing route to speedtest.xilo.net [91.199.78.82]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms t1 [10.1.0.21]
2 <1 ms <1 ms <1 ms gateway.lan [10.0.0.1]
3 6 ms 6 ms 6 ms gateway.bt [217.32.142.104]
4 6 ms 6 ms 6 ms 217.32.142.158
5 14 ms 14 ms 14 ms 213.120.163.90
6 13 ms 14 ms 13 ms 217.32.27.82
7 14 ms 14 ms 14 ms 217.32.27.178
8 14 ms 14 ms 13 ms acc1-10GigE-7-2-0.mr.21cn-ipp.bt.net [109.159.250.96]
9 24 ms 23 ms 23 ms core1-te0-12-0-1.ealing.ukcore.bt.net [109.159.250.22]
10 20 ms 20 ms 20 ms peer2-xe0-0-0.redbus.ukcore.bt.net [109.159.254.74]
11 20 ms 20 ms 20 ms 195.99.126.103
12 21 ms 20 ms 20 ms uk-dedicated-servers-154.gw.goscomb.net [93.89.90.154]
13 20 ms 20 ms 20 ms the-tcx.ukservers.com [78.110.166.101]
14 21 ms 21 ms 21 ms ge-0-1-0-6.v-243.rtr-1.bs3.mhd.uk.as44574.net [78.110.166.90]
15 23 ms 22 ms 22 ms ge-0-0-0.fw-c-0.bs3.mhd.uk.as44574.net [91.199.78.1]
16 23 ms 22 ms 23 ms speedtest.xilo.net [91.199.78.82]

Getting from London to Manchester in 2ms isn't possible on any commercial network, so it cannot be in Manchester. It was at the time within 2ms of London.

Since I reported it the server pointed to by the Manchester (Xilo) speed test has changed, it's now 31.193.12.35 so I guess they fixed it without responding to my support ticket.

Edited by deleted (Tue 08-May-12 23:13:54)

In reply to a post by RobFlet:

my IPP has dropped to 57/20 from 66/20 over the last couple of weeks.

You could well be on an interleaved connection, then.

In reply to a post by RobFlet:

On the 80/20, I am seeing 55/13 over my homeplugs - so can't complain - but I have had one instance of a slowdown to 0.5megs until I reconnected the HH3.

Are you using the Home-plugs when getting the poor (ish) pings?
Are they any better if directly connected via Ethernet?

In reply to a post by RobFlet:

Good, I'm glad it hasn't moved ...

Mind you, when I was on Infinity 40/10, I was getting pings to Manchester of around 10-15 ms ...

Now on 80/20, I get pings around 30+ to Manchester ...

Any ideas???

Rob

Well actually it has.

Or rather, it now points to a different sever, under the same label.

Am I missing something?
Where on that traceroute does it indicate a 2ms ping from Manchester to London?
I see no 2ms ping anywhere on the traceroute.

In reply to a post by adebov:

Am I missing something?
Where on that traceroute does it indicate a 2ms ping from Manchester to London?
I see no 2ms ping anywhere on the traceroute.

C:\Users\Administrator>tracert 91.199.78.82

Tracing route to speedtest.xilo.net [91.199.78.82]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms t1 [10.1.0.21]
2 <1 ms <1 ms <1 ms gateway.lan [10.0.0.1]
3 6 ms 6 ms 6 ms gateway.bt [217.32.142.104]
4 6 ms 6 ms 6 ms 217.32.142.158
5 14 ms 14 ms 14 ms 213.120.163.90
6 13 ms 14 ms 13 ms 217.32.27.82
7 14 ms 14 ms 14 ms 217.32.27.178
8 14 ms 14 ms 13 ms acc1-10GigE-7-2-0.mr.21cn-ipp.bt.net [109.159.250.96]
9 24 ms 23 ms 23 ms core1-te0-12-0-1.ealing.ukcore.bt.net [109.159.250.22]
10 20 ms 20 ms 20 ms peer2-xe0-0-0.redbus.ukcore.bt.net [109.159.254.74]
11 20 ms 20 ms 20 ms 195.99.126.103
12 21 ms 20 ms 20 ms uk-dedicated-servers-154.gw.goscomb.net [93.89.90.154]
13 20 ms 20 ms 20 ms the-tcx.ukservers.com [78.110.166.101] <== London = 20ms
14 21 ms 21 ms 21 ms ge-0-1-0-6.v-243.rtr-1.bs3.mhd.uk.as44574.net [78.110.166.90]
15 23 ms 22 ms 22 ms ge-0-0-0.fw-c-0.bs3.mhd.uk.as44574.net [91.199.78.1]
16 23 ms 22 ms 23 ms speedtest.xilo.net [91.199.78.82] <== Destination = 22ms

22 - 20 = 2ms.

In reply to a post by qasdfdsaq:

22 - 20 = 2ms.

Isn't each set of ping figures the ping between you and the server at that step on the traceroute? And the ping figures are not really related to how long it takes to get from one step to the next (and are more related to how long each server takes to get a message back to you). EDIT: Or how long each server takes to get round to responding to pings (based on the fact that's not it's primary function - pings often take a lower priority to actually routing).
Using your logic it took minus 4ms to get a message from Ealing to London and that just doesn't work.

Edited by adebov (Tue 08-May-12 23:29:27)

In reply to a post by adebov:

In reply to a post by qasdfdsaq:

22 - 20 = 2ms.

Isn't each set of ping figures the ping between you and the server at that step on the traceroute? And the ping figures are not really related to how long it takes to get from one step to the next (and are more related to how long each server takes to get a message back to you).

If it takes 20ms to get from A to B, and 22ms to get from A to B to C then the difference is the time from B to C. Simple logic.

And no, traceroutes don't talk to "servers" nor do they "ping" any steps along the way, in fact Linux traceroutes usually don't ping anything at all.

If you want to learn about how traceroutes work, try here.

Using your logic it took minus 4ms to get a message from Ealing to London and that just doesn't work.

One router that does not respond to ICMP in a timely manner does not change the conclusion.

Edited by deleted (Tue 08-May-12 23:38:12)