I switched from BeThere yesterday and am moderately disappointed ping wise. I'm in SW London and was getting a 5ms first hop with 5-7ms to most London based servers.

Now I get a 5ms first hop with FTTC, but it jumps around for 15ms before reaching LINX - so I suspect would get 20-22ms to the IPs below. It's pretty baffling to be honest, where on earth are they routing things?

I'll post some tracerts tonight when I get home from work.

FWIW, I am/was a very keen QW player, so understand the importance of latency in what looks like QuakeLive server IPs?

Yes I am a QuakeLive player, and yeah I fully understand the importance of latency, at the moment I get about 14ms to the first, 19ms to the 2nd and 30ms to the last, every 10ms makes a difference in this game, it starts becoming unplayable for me when its over 50ms so really im looking for maximum 45ms to German servers (as they are the most populated servers in this game).

BTW im from NW England so I guess these are very good pings and it looks unlikely that I would get close to that on BT FTTC, 14ms to most UK servers is what I currently get. On BT it would probably be +15ms to everything I currently have?

24
23
38
Infinity2 N. Ireland

How can you be disappointed with a 5-20ms ping? That's incredibly low and much lower than you need for professional level gaming. If you're not performing well with a 20ms ping then I think you should look at your own gaming skills rather than blaming it on latency 20ms is so small as to be insignificant and I would doubt anyone could tell the difference between 20ms and 5ms as the gap is so incredibly tiny (and insignificant compared to the human reaction time).

There's a few reasons I'm a bit disappointed:

- My ping to London based servers is now 4 times higher than it was before for no apparent purpose. It used to be 5ms with Be and ADSL, now it is 20ms with BT and VDSL2. I'm less than 15 miles from LINX and Telecity, with a first hop of 5ms. How do they manage to add 15ms in 15 miles aside from nonsensical routing? It seems very very inefficient.

- Most QW gaming happens on continental servers - particularly denmark. I use to have about 22-24ms to qw.foppa.dk - this is now 31ms, for no apparent reason.

- Further to the above, QW uses 77hz as its server tick frequency, which roughly correlates to 12ms/25ms/38ms for each frame. Your in game ping will 'snap' to one of these values. So my danish ping used to be 25ms (ie, 2 frames delay) and is now 38ms (ie 3 frames delay). The difference between 12ms and 25ms is minimal. The difference between 25ms and 38ms is noticeable, but OK. 52ms is very noticeable - above this is not so fun.

- Having a bonus 15ms for no apparent reason means I'm more likely to have to play with a ping of 52ms or more on continental servers.

- It's just plain inefficient!

It's not like 20ms is _bad_ per se and it's completely playable, it's just 4 times higher than it needs to be especially considering the distance. And if for some reason in the future interleaving has to be automagically added to my line I'm probably looking at 30ms for UK servers minimum - which I do actually consider quite bad to servers in the same city as me.

I don't really know enough about BT's network to say whether you'd definitely see another 15ms - there are people here posting from much further away than me with the same or lower pings (eg 22ms from Northern Ireland I think, 19ms from Edinburgh). Maybe it's fine up north as there might be fewer opportunities for them to do dumb things on the way south?

Edited by deleted (Tue 29-May-12 12:22:17)

I believe BT has on the order of a dozen BRAS dial-in sites around the country. When your router connects to "the internet", it dials into one of these BRAS "servers" and your traffic gets routed from there, to your ISP's core network and then typically down to London.

Your router would normally pick the first BRAS to respond when it sends out a request. This may or may not be the one nearest you. However you could theoretically (within some limits, which I'm not particularly familiar with) connect to a different BRAS, either by chance or because your primary is down. In my case, the nearest alternatives would be Glasgow, Newcastle or Manchester, none of which are really "further" than Edinburgh anyway, so yes, less opportunity for dumb things to happen.

If you're in London though, and by chance connect to the Newcastle BRAS, your traffic would go London => Newcastle => <various other intermediate hops> => London. That would be stupid. But that could easily be diagnosed on a traceroute, a distant BRAS would show up as your first hop (default gateway) being high, which doesn't seem to be the case for driz. That could also happen with interleaving, which you'd need a hacked modem to rule out.

I'd be interested to see what the traceroutes show.

For the record, on Be (well, old O2 legacy package) I was getting 26ms to London servers and now on BT Infinity I'm getting 18-20ms to the same servers, on the same line. Both connections with interleaving off. VM were managing 16ms minimum ping but an average of 25-30 thanks to record-breaking jitter...

I've added a few of the random hops at the beginning to my hosts file to try and keep track of what these IPs are and when they show up.

Here's a tracert to bbc.co.uk:

Tracing route to bbc.co.uk [212.58.241.131]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms a.router [10.1.1.254]
2 6 ms 6 ms 6 ms slowhop2 [217.32.147.97]
3 6 ms 6 ms 6 ms randhop3 [217.32.147.142]
4 12 ms 11 ms 11 ms randhop4 [213.120.181.142]
5 11 ms 11 ms 11 ms slowhop5 [217.41.169.197]
6 12 ms 11 ms 11 ms slowhop6 [217.41.169.109]
7 11 ms 11 ms 11 ms 109.159.251.235
8 22 ms 27 ms 24 ms core2-te0-13-0-2.ealing.ukcore.bt.net [109.159.251.139]
9 16 ms 22 ms 83 ms peer2-xe10-1-0.telehouse.ukcore.bt.net [109.159.254.116]
10 19 ms 18 ms 17 ms 194.74.65.42
11 * * * Request timed out.
12 22 ms 28 ms 22 ms ae1.er01.rbsov.bbc.co.uk [132.185.254.46]
13 18 ms 19 ms 21 ms 132.185.255.134
14 18 ms 20 ms 19 ms 212.58.241.131

And to the first server in this post:

Tracing route to quakelive.london.i3d.net [94.76.229.13]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms a.router [10.1.1.254]
2 6 ms 5 ms 6 ms slowhop2 [217.32.147.97]
3 7 ms 8 ms 23 ms randhop3 [217.32.147.142]
4 11 ms 11 ms 11 ms 213.120.181.126
5 11 ms 11 ms 11 ms slowhop5 [217.41.169.197]
6 11 ms 11 ms 11 ms slowhop6 [217.41.169.109]
7 11 ms 11 ms 11 ms 109.159.251.231
8 20 ms 19 ms 19 ms core1-te0-12-0-6.ilford.ukcore.bt.net [109.159.251.133]
9 26 ms 23 ms 20 ms peer4te0-1-0-0.telehouse.ukcore.bt.net [109.159.254.54]
10 18 ms 18 ms 18 ms hex89-linx.as29550.net [195.66.236.223]
11 22 ms 20 ms 19 ms vl668.the-7600.as29550.net [92.48.95.10]
12 21 ms 21 ms 26 ms a.1.magic-the.as29550.net [217.112.81.90]
13 18 ms 18 ms 17 ms quakelive.london.i3d.net [94.76.229.13]

Compared with my old Be connect to jolt.co.uk:

Tracing route to jolt.co.uk [84.234.17.86]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms billion.router [192.168.1.254]
2 * * * Request timed out.
3 * 6 ms 6 ms 10.1.2.218
4 6 ms 6 ms 5 ms linx-extreme1.netrino.co.uk [195.66.227.18]
5 6 ms 5 ms 6 ms thn-edge.netrino.co.uk [84.234.19.142]
6 6 ms 5 ms 5 ms hex-edge.netrino.co.uk [84.234.19.146]
7 7 ms 7 ms 6 ms bs1-core.netrino.co.uk [84.234.19.5]
8 7 ms 7 ms 6 ms bruce.ferrago.net [84.234.17.86]

Going to try rebooting the router/restarting the PPPoE session to see what happens.