BT Infinity Network Infrastructure - The reason behind BT being the worst UK ISP for gaming

Associate
Joined
26 Jul 2019
Posts
6
As a competitive gamer, by that I mean someone who has competed for cash prizes at events, rather than just a hardcore player, i'm constantly on the lookout for subtle changes in my network which can cause major issues.

After playing and competing Fifa Ultimate team for a number of years, one thing was always clear, the servers of the game in general are not optimally set....However, if like me you're getting your internet through BT, specifically BT infinity, the major problem is likely to be the way in which BT's infrastructure is so bizarrely maintained.


So, I did a little digging with the help of a friend who uses Sky broadband.
By both running a traceroute through the address provided by EA, it is clear that the major problem with connectivity/delay/lag is coming from BT.

BT Tracert

1 10.0.0.1 1 1ms
2 * Unknown Host * 0.0.0.0 2 0ms
3 * Unknown Host * 31.55.187.181 3 14ms
4 * Unknown Host * 31.55.187.180 4 13ms
5 core2-hu0-17-0-5.southbank.ukcore195.99.127.190 5 14ms
6 peer2-et0-1-5.slough.ukcore.bt.ne62.172.103.210 6 13ms
7 * Unknown Host * 195.66.224.116 7 14ms
8 * Unknown Host * 0.0.0.0 8 0ms
9 * Unknown Host * 134.213.97.149 9 16ms
10 corea-core6.lon3.rackspace.net 164.177.137.13 10 16ms
11 core6-eaaggr1a.lon3.rackspace.net31.222.133.213 11 17ms
12 * Unknown Host * 134.213.37.197 12 14ms



Sky Tracert

1 <1 ms 192.168.1.1
2 5 ms 97e7fef2.skybroadband.com [151.231.254.242]
3 10 ms be259.pr2.hobir.isp.sky.com [89.200.131.2]
4 10 ms rackspace-ip81-27.no-dns-yet.lonap.net [5.57.81.27]
5 * * * Request timed out.
6 13 ms 134.213.97.153
7 14 ms coreb-core6.lon3.rackspace.net [164.177.137.25]
8 12 ms area101.core2-core1.lon.rackspace.net [31.222.133.227]
9 12 ms 134.213.37.197



On the face of it, the ping times are very good.
The average gamer would likely not even notice.
But what becomes a noticeable issue, is the vast number of hops it takes to leave both the BT and Sky internal network, as the final 4 hops on both networks are owned by EA and actually are practically identical.


On the BT route -

3 * Unknown Host * 31.55.187.181 3 14ms
4 * Unknown Host * 31.55.187.180 4 13ms
5 core2-hu0-17-0-5.southbank.ukcore195.99.127.190 5 14ms
6 peer2-et0-1-5.slough.ukcore.bt.ne62.172.103.210 6 13ms


On the Sky route -

2 5 ms 97e7fef2.skybroadband.com [151.231.254.242]
3 10 ms be259.pr2.hobir.isp.sky.com [89.200.131.2]



For some strange reason, for which I am unable to get a response from BT about, there appears to be some kind of what I would call 'Dual routing' going on.
For some reason the infrastructure has twice as many hops, on practically identical servers.

It isn't some kind of one of either. Looking at the route to a server outside of the UK which hosts dedicated servers for games such as Call of duty.....

Host Name IP Address Hop Ping Time
10.0.0.1 1 2ms
* Unknown Host * 0.0.0.0 2 0ms
* Unknown Host * 31.55.187.177 3 11ms
* Unknown Host * 31.55.187.176 4 15ms
core1-hu0-12-0-3.southbank.ukcore195.99.127.44 5 12ms
peer6-hu0-6-0-6.telehouse.ukcore.195.99.127.3 6 14ms
t2c3-et-3-3-0-0.uk-lon1.eu.bt.net166.49.211.238 7 11ms
166-49-211-38.eu.bt.net 166.49.211.38 8 22ms


Once again, Dual hops at soon as it leaves my internal network, when it reaches the London main server, and when it joins the BT server to leave the UK.



This is why BT is so poor compared to other UK providers.

Twice as many hops = Twice as much latency + Twice as much opportunity for packet loss.
 
Associate
OP
Joined
26 Jul 2019
Posts
6
They don't advertise as a gaming network though that's the thing. Gamers should be on something better than a BT connection more dedicated for throughput/hops e.t.c.

So you're saying, that if you only had the choice between the two above, using the tracert provided, Sky is the obvious choice?
 
Associate
OP
Joined
26 Jul 2019
Posts
6
The lag etc might be coming from other players though and without knowing what route they take it's hard to tell. Unless all the gamers are within the UK? Do any of the ISPs traffic shape for gaming?

The game is hosted on dedicated servers.

The above two traceroutes are which BT and Sky take from my area to the dedicated servers in question.

Sky and Virgin both take virtually the same 2 hops to get out of their network, suggesting that their network is shaped pretty well.

Whereas BT always take twice the hops.
 
Associate
Joined
6 Jun 2016
Posts
1,569

1 <1 ms <1 ms <1 ms *.*.*.*
2 <1 ms <1 ms <1 ms *.*.*.*
3 3 ms 2 ms 2 ms manc-lam-4-tenge13-3500.network.virginmedia.net [82.15.205.18]
4 2 ms 2 ms 2 ms manc-core-2a-xe-111-0.network.virginmedia.net [213.104.242.25]
5 * * * Request timed out.
6 9 ms 9 ms 9 ms 86.85-254-62.static.virginmediabusiness.co.uk [62.254.85.86]
7 9 ms 9 ms 9 ms 213.46.175.250
8 9 ms 9 ms 9 ms ae13.mpr3.lhr3.uk.zip.zayo.com [64.125.30.54]
9 10 ms 10 ms 10 ms ae14.mpr1.lhr23.uk.zip.zayo.com [64.125.30.173]
10 11 ms 10 ms 10 ms 94.31.42.254.ipyx-076520-zyo.above.net [94.31.42.254]
11 * * * Request timed out.
12 12 ms 12 ms 12 ms 134.213.131.251
13 12 ms 12 ms 12 ms corec-coreb.lon5.rackspace.net [162.13.232.14]
14 13 ms 12 ms 12 ms coreb-core6.lon3.rackspace.net [164.177.137.25]
15 12 ms 12 ms 12 ms aggr331b-4-aggr331a-4.lon3.rackspace.net [31.222.133.163]
16 12 ms 12 ms 12 ms 134.213.37.197
 
Associate
OP
Joined
26 Jul 2019
Posts
6
1 <1 ms <1 ms <1 ms *.*.*.*
2 <1 ms <1 ms <1 ms *.*.*.*
3 3 ms 2 ms 2 ms manc-lam-4-tenge13-3500.network.virginmedia.net [82.15.205.18]
4 2 ms 2 ms 2 ms manc-core-2a-xe-111-0.network.virginmedia.net [213.104.242.25]
5 * * * Request timed out.
6 9 ms 9 ms 9 ms 86.85-254-62.static.virginmediabusiness.co.uk [62.254.85.86]
7 9 ms 9 ms 9 ms 213.46.175.250
8 9 ms 9 ms 9 ms ae13.mpr3.lhr3.uk.zip.zayo.com [64.125.30.54]
9 10 ms 10 ms 10 ms ae14.mpr1.lhr23.uk.zip.zayo.com [64.125.30.173]
10 11 ms 10 ms 10 ms 94.31.42.254.ipyx-076520-zyo.above.net [94.31.42.254]
11 * * * Request timed out.
12 12 ms 12 ms 12 ms 134.213.131.251
13 12 ms 12 ms 12 ms corec-coreb.lon5.rackspace.net [162.13.232.14]
14 13 ms 12 ms 12 ms coreb-core6.lon3.rackspace.net [164.177.137.25]
15 12 ms 12 ms 12 ms aggr331b-4-aggr331a-4.lon3.rackspace.net [31.222.133.163]
16 12 ms 12 ms 12 ms 134.213.37.197

Are you using two routers?

I've seen numerous people have issues with the Zayo servers aswell especially with packet loss.
 
Associate
Joined
23 Feb 2009
Posts
2,396
Location
Bournemouth
Here is mine on zen.

We could do a comparison between me on zen, you on bt and your friend on sky.

C:\Users\Dan>tracert 134.213.37.197

Tracing route to 134.213.37.197 over a maximum of 30 hops

1 1 ms 2 ms 1 ms fritz.box [192.168.178.1]
2 20 ms 38 ms 14 ms vt1.cor2.lond1.ptn.zen.net.uk [51.148.72.22]
3 20 ms 16 ms 15 ms ae-7.cor2.lond2.ptn.zen.net.uk [51.148.73.13]
4 15 ms 14 ms 16 ms ae-21.agg3.lond2.ptn.zen.net.uk [51.148.73.40]
5 16 ms 28 ms 134 ms 195.66.224.116
6 * * * Request timed out.
7 17 ms 17 ms 17 ms 134.213.97.211
8 18 ms 18 ms 17 ms cored-coreb.lon5.rackspace.net [162.13.232.18]
9 23 ms 18 ms 18 ms coreb-core6.lon3.rackspace.net [164.177.137.25]
10 20 ms 17 ms 17 ms core6-eaaggr1b.lon3.rackspace.net [31.222.133.219]
11 17 ms 17 ms 17 ms 134.213.37.197
 
Caporegime
Joined
17 Jul 2010
Posts
25,767
Sky Superfast 80/20 FTTC:

1 <1 ms <1 ms <1 ms SkyRouter.Home [192.168.0.1]
2 4 ms 4 ms 4 ms 97e7fedc.skybroadband.com [151.231.254.220]
3 19 ms 14 ms 14 ms be403.pr2.enlba.isp.sky.com [2.120.9.106]
4 14 ms 17 ms 14 ms rackspace-ip81-27.no-dns-yet.lonap.net [5.57.81.27]
5 * * * Request timed out.
6 19 ms 17 ms 16 ms 134.213.97.149
7 17 ms 16 ms 16 ms corea-core6.lon3.rackspace.net [164.177.137.13]
8 17 ms 17 ms 17 ms core6-eaaggr1a.lon3.rackspace.net [31.222.133.213]
9 23 ms 14 ms 15 ms 134.213.37.197

Running Apex, uTorrent and Discord.
 
Soldato
Joined
5 Nov 2011
Posts
5,368
Location
Derbyshire
FWIW here's a PlusNet trace (pardon my Linux). My line is pretty bad ATM and I have a fault raised due to excessive errors and underperforming sync speeds.
Code:
general (10.10.10.34)                                                                          2019-07-26T23:35:01+0100
Keys:  Help   Display mode   Restart statistics   Order of fields   quit
                                                                               Packets               Pings
 Host                                                                        Loss%   Snt   Last   Avg  Best  Wrst StDev
 1. router                                                                    0.0%     7    0.4   0.4   0.3   0.6   0.1
 2. 84.93.253.88                                                              0.0%     7   20.2  20.2  19.7  20.9   0.4
 3. 84.93.253.95                                                              0.0%     7   20.5  20.7  20.5  21.2   0.3
 4. 195.99.125.144                                                            0.0%     7   20.7  20.8  20.2  21.2   0.3
 5. peer2-et-3-0-4.slough.ukcore.bt.net                                       0.0%     7   21.2  21.8  21.1  25.1   1.5
 6. ???
 7. ???
 8. 134.213.131.251                                                           0.0%     6   22.8  22.9  22.8  23.0   0.1
 9. corec-coreb.lon5.rackspace.net                                            0.0%     6   24.5  24.6  24.4  24.9   0.2
10. coreb-core6.lon3.rackspace.net                                            0.0%     6   23.8  24.0  23.8  24.4   0.3
11. core6-eaaggr1a.lon3.rackspace.net                                         0.0%     6   23.8  23.8  23.6  24.0   0.2
12. 134.213.37.197                                                            0.0%     6   23.6  23.5  23.2  23.9   0.2

The long and short of this is that if you are going to make money off it or are dependant on it. Pay for a REAL service with SLA's and such.
 
Caporegime
Joined
18 Oct 2002
Posts
26,116
You have absolutely no idea how many actual hops there are on that path, and you're drawing a conclusion based on faulty logic.

Here's a leased line:
Code:
gateway # execute traceroute 134.213.37.197
traceroute to 134.213.37.197 (134.213.37.197), 32 hops max, 3 probe packets per hop, 72 byte packets
 1  * <*.exponential-e.net>  98.851 ms  6.437 ms  9.935 ms
 2  * 195.66.224.186 <1-1-2.pr01.thca.uk.exponential-e.net>  4.685 ms  4.645 ms
 3  195.66.224.116  5.836 ms  6.144 ms  6.123 ms
 4  * * *
 5  134.213.97.213  7.443 ms  7.318 ms  7.304 ms
 6  164.177.137.25 <coreb-core6.lon3.rackspace.net>  6.227 ms  6.156 ms  6.243 ms
 7  31.222.133.219 <core6-eaaggr1b.lon3.rackspace.net>  5.965 ms  5.983 ms  5.949 ms
 8  134.213.37.197  6.638 ms  6.644 ms  6.668 ms


I know for a fact that there's a lot of network before hop 1 on that traceroute, because it's 100 miles away. But it's hidden inside the service provider's network.
 
Last edited:
Caporegime
Joined
18 Oct 2002
Posts
26,116
You could have 50 hops inside an MPLS path and traceroute would only show you 2.

For example if you VPN into your work or back to your home and do a traceroute to a PC on that network you will see two or three hops, but that clearly isn't possible.
 
Soldato
Joined
30 Jul 2007
Posts
5,185
Location
Lincolnshire
Yeah don't really think you can complain on a residential product.

To compare, Plusnet 80/20 FTTC:

Code:
 1    <1 ms    <1 ms    <1 ms  BTHUB [192.168.1.1]
  2    10 ms    12 ms    10 ms  84.93.253.88
  3    11 ms    11 ms    11 ms  84.93.253.91
  4    10 ms    10 ms    11 ms  core1-be1.colindale.ukcore.bt.net [195.99.125.132]
  5    11 ms    10 ms    10 ms  peer2-et3-0-1.slough.ukcore.bt.net [62.172.103.214]
  6     *        *        *     Request timed out.
  7     *        *        *     Request timed out.
  8    13 ms    12 ms    12 ms  134.213.97.213
  9    22 ms    13 ms    13 ms  coreb-core6.lon3.rackspace.net [164.177.137.25]
 10    13 ms    13 ms    13 ms  area101.core2-core1.lon.rackspace.net [31.222.133.227]
 11    11 ms    11 ms    11 ms  134.213.37.197
 
Permabanned
Joined
9 Aug 2008
Posts
35,707
You could have 50 hops inside an MPLS path and traceroute would only show you 2.

For example if you VPN into your work or back to your home and do a traceroute to a PC on that network you will see two or three hops, but that clearly isn't possible.

So traceroute is pretty redundant then and shouldn't be used to detect network problems like this? Then what other tools are accurate in doing so?
 
Caporegime
Joined
18 Oct 2002
Posts
26,116
It's still a useful tool, but you can't really look at the hop count between two different ISPs and declare one to be better for gaming. Network troubleshooting is very difficult if you don't have access to the devices that make that network up, and an idea of how they are all connected together.
 
Back
Top Bottom