Thanks, I'm still looking in to it but I'm on leave today I'm pretty sure it's not a general problem and not latency you'd see in normal use. We don't see the spikes on the graphs we produce ourselves to customer lines. - I think the spikes are caused by something between the ThinkBroadband hosting provider and us.

I changed the routing between TBB and AA at 9PM on 6th June - see if that makes a difference. It's now going via LONAP instead of LINX. If that makes a difference we'll know where to look to get to the bottom of it.

Looking great on IPv4 this morning.

https://www.thinkbroadband.com/broadband/monitoring/...

Should IPv6 show improvements yet?

https://www.thinkbroadband.com/broadband/monitoring/...

Whilst I've never seen any issues with the service despite the BQMs on TBB, it would be great to see some flat graphs so that should there be some problem it would show up more clearly.

Just a quick addition, assuming the route change is IPv4 only so far, this seems to have had a negative effect on the TBB speed tester, with throughput on IPv4 single thread dropping quite a bit for me, it has always been on par with IPv6 when testing on TBB. If I do a route trace I can see the speedtest server is also via lonap-gw1. You'll wish you never started this

IPv6
https://www.thinkbroadband.com/speedtest/16545873841...

IPv4
https://www.thinkbroadband.com/speedtest/16545874576...

Edited by E300 (Tue 07-Jun-22 08:47:44)

IPv4 is much better though still not brilliant. Negligible improvement, if any, on IPv6.

Both IPv4 and IPv6 speed tests here are excellent, single and multi-threads.

Sorry - I meant to say that I only changed routing for IPv4 - IPv6 is still routing via LINX as before for comparison.

What speed package are you on? For me it seems IPv4 hits a ceiling on single threaded tests to TBB of around 350Mbps since the routing change. I don't have any speed tests immediately prior to the change but looking back at the last few I did a few weeks ago IPv4 was >650 on single thread tests. Just tested on Speedtest.net on their single thread option and > 750.

Obviously this is just affecting a speed test and everything else is great, just thought an interesting observation and indication of how a different routing can give different results.

As for the BQM on IPv4 for me now mirrors what I've seen with other ISPs prior to joining AAISP, and its the best it could be now. Can you link your BQMs so we can take a look.

Edited by E300 (Tue 07-Jun-22 11:27:14)

I'm only on FTTC 80/20.

In reply to a post by E300:

As for the BQM on IPv4 for me now mirrors what I've seen with other ISPs prior to joining AAISP, and its the best it could be now. Can you link your BQMs so we can take a look.

Here's my IPv4 BQM

https://www.thinkbroadband.com/broadband/monitoring/...

Ahh, that doesn't look great, perhaps you have an extra issue. Always worth setting up an alternative BQM, you can set up a free one at https://f8lure.mouselike.org/, this also sets up one from AAISP as well automatically, so you get two for one. See if that shows the same issues. The f8lure charts for me have never shown the spikes seen on TTB BQM.

Also may be worth posting a tracert to the TBB one, just to confirm you are routing the same way, mine is:

Text

1 23 45 67 89 1011

Tracing route to pingbox1.thinkbroadband.com [80.249.99.164] over a maximum of 30 hops:    1    <1 ms    <1 ms    <1 ms  pfSense.localdomain [192.168.1.1]  2     7 ms     7 ms     7 ms  y.witless.thn.aa.net.uk [90.155.53.134]   3     7 ms     7 ms     7 ms  q-aimless.tch.aa.net.uk [90.155.53.109]  4     7 ms     7 ms     7 ms  lonap-gw1.thdo.ncuk.net [5.57.80.142]   5     7 ms     7 ms     7 ms  po4-31.core-rs4.thdo.ncuk.net [80.249.97.85]  6     7 ms     7 ms     7 ms  pingbox1.thinkbroadband.com [80.249.99.164]  Trace complete.