The speedtest server is on ookla.btlancashire.co.uk [194.150.176.123], but same trace route.

From a quick Google it would look like BT Lancashire Services is a services company owned by BT that provide IT and payment services etc for Lancashire County Council, which is also the organisation name on the whois record.

Don't JANET offer "govroam" to public bodies? So I wouldn't be surprised if they also offer connectivity to public bodies too.

Although it does sound an odd setup!!

Almost as odd as Vodafone Broadband at the moment!

Hum, I pretty sure that they are not supposed to be doing that. It certainly looks like BT Lancashire Services have not thought about whether they should be doing that. I will try and catch our head of networks and see what he as to say.

Mind you the JANET backbone is multiple 400Gbps links all over the shop with 1.2Tbps of external connectivity last I heard so one would think it should be well enough connected to be a speedtest server.

I have just done a speedtest from a remote desktop on my 10Gbps connected console server at work against the BT Lancashire Services speed test server and I would say from the results it would only appear to have a 1Gbps connection

Ping ms 14
Download Mbps 816.41
Upload Mbps 622.18

I know my server can go much faster because I have tested it against a machine at another UK university and the traceroute shows that I never step outside JANET to get to it either. Mind you that's better than a thinkbroadband speed test from the same server.

Of course it's not really fair testing them with a 10Gbps connection

It's not only odd, I am pretty certain they should not be doing it.

That is the service company for Lancashire County Council should not be running a speedtest server for BT plc over the JANET network, even the service company is owned by BT.

It's basically a misuse of public sector money meant for research and teaching by a public limited company.

It does look that way but it could be there for some specific reason and I doubt the site itself gives full visibility but I think the wider context of the test itself isn't proving that tester has any advantage over the BT Wholesale one.

Maybe a misunderstanding from someone at Vodafone that a "BT" test should be used and it was one they found. Who knows.

Matt

Yeah there seems to be some "misunderstanding" at Vodafone at the moment IMHO!!

I think the OP's point really is that VF are asking people not to use their (speedtest.vodafone.co.uk) server when making complaints about unsatisfactory performance as it is giving poor results and they say it has issues. As you say probably due to misunderstanding but coincidentally the route to the BTLancashire server seems to give "good" speed/results.

The VF server gives me very poor results when on my VF connection at home (24/7). Amusingly if I speedtest using the server from work on a Virgin Business connection it gives results in line with other servers tested and not bad results if I test it through a VPN on my home VF connection, so the server itself is probably OK but routing from VF broadband to it probably isn't.

I appreciate congestion and all the analogies to motorways and traffic jams, as well as various routing, traffic segregation and CDNs for some services, but I can achieve good speed throughput on some routes 24/7, others are very poor 24/7 plus some overall congestion at peak which is to be expected.

This all started fairly suddenly in Nov 18, I wonder what experience people on the CityFibre/VF gigafast are having or if anyone knows if they use different core routing? If its the same and they get good takeup it doesn't bode well.

Another "it was good up till November" person. It seems the sooner the routing leaves VF's network, the better the speed tests.

Even the BT Lancs server was struggling last night, so I took a few screenshots to sent to VF (it got as slow as 25Mbps).

It wouldn't be so bad if VF accepted there were issues and that they were doing address them, but instead they deny there are any problems.

My SamKnows box shows a drop off in NetFlix performance around the same time.

Yep, gained my details from here:

https://www.speedtestserver.com/

Country: Great Britain
City: Preston
Provider: BT Lancashire Services
Host/Port: ookla.btlancashire.co.uk:8080
Speedtest ID: 9229

C:\Users\matth>tracert ookla.btlancashire.co.uk

Tracing route to ookla.btlancashire.co.uk [194.150.176.123]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms 192.168.100.254
2 1 ms 1 ms 1 ms 192.168.1.1
3 9 ms 8 ms 10 ms host-212-158-250-33.dslgb.com [212.158.250.33]
4 10 ms 10 ms 10 ms 63.130.105.130
5 11 ms 11 ms 11 ms linx-gw1.ja.net [195.66.224.15]
6 12 ms 12 ms 12 ms ae23.londtt-sbr1.ja.net [146.97.35.169]
7 16 ms 15 ms 16 ms ae27.erdiss-sbr2.ja.net [146.97.33.14]
8 14 ms 14 ms 14 ms ae29.manckh-sbr2.ja.net [146.97.33.42]
9 15 ms 14 ms 14 ms ae25.manckh-ban1.ja.net [146.97.35.50]
10 14 ms 14 ms 14 ms lancs-council.manckh-ban1.ja.net [146.97.40.50]
11 * * * Request timed out.
12 * ^C
C:\Users\matth>

Probably a few of us giving it some stick, or you were on at the same time as jabuzzard testing from a 10Gbps link!!!

Yeah FreshMint's graph over on the VF forum thread I see you've commented on is rather a clear demonstration.

Their lack of acknowledgement to customers is a shame. The dramatic change and the fact effected routes/traffic are slow 24/7 (It doesn't sit right that routes are congested 24/7 unless a crazy routing error/provision) leads me to wonder if this is a commercial decision they have taken.

Edited by adrenalize_ (Fri 01-Feb-19 15:47:33)

Maybe Dilbert has the answer: Dilbert: Internet is Slow