I didn't even know about the line data access!

I've just had a look and it seems the data indicates that my line speed has been identical for the past month (upstream: 19999, downstream: 79911, sync rate: 79911).

I download and upload large files frequently, and sometimes find myself noticing the difference in speed compared to when I used Virgin Media (on the few occasions it worked...). That's probably why I became fixated on my speed test, when it reality it probably won't make any difference if I did manage to eke out those last few Mbps.

You're absolutely right in that that Zen are great, especially compared to Virgin Media, in every aspect. Zen is far superior when it comes to gaming too.

But I do wonder why ISPs don't simply advertise a real-world minimum guaranteed download/upload speed for a given address rather than go by sync speeds, taking into consideration the known protocol overheads.

I can see how that would be challenging given so many variables to account for, but I'm sure it's possible. Especially when there's been such an uproar about advertised internet speeds for so long. As a consumer, I would be chuffed if my speed tests came out better than my minimum guaranteed speeds (shockingly, Virgin Media actually seem to do this).

I would venture that many consumers don't know about sync speeds and protocol overheads (I didn't until I started researching here, none of my friends and family did either when asked).

It feels a bit like buying a pizza but one of the slices is inedible because it's stuck to the lid.

I so wish I had FTTP coming to my area, but that appears to be a few years away at the very least.

Well, indeed speed tests vary considerably. I get around 75-76Mbps with Speedtest.net using the Windows app. With broadbandspeedchecker.co.uk I get 77-78Mbps speedguide.net another site I get 80Mbps.

Of-course the Speedofme is exaggerated. There's no way we are getting 82Mbps. But my point was that even with those speed test sites that I mention where I get 80/20 speed test results, these you'll not get if your router isn't syncing at the full speeds. But AAISP speed test shows

"This speed tester is hosted on the same network that broadband connections are terminated on, so this should be the closest speed tester for AAISP customers." It will apply for the rest of the ISPs on FTTC as they are on the same Openreach network.

Originally, I had the impression that you were only syncing at 73Mbps max in your router stats and not just speed tests giving you that number. That would've been a cause for concern given that your line is only 118 meters long.

But it turns out that indeed you are syncing at full sync speeds in your Fritz!Box and that for me is ultimately the most important thing. We aren't always going to get 80Mbps each time even if it was accurate. But it usually is a positive indicator that your router is syncing at full sync speeds.

Also the browser is very important. With Chromium browsers such as Opera and Opera GX Gaming Browser I get usually much better speeds and lower pings compared to Waterfox and Firefox web browsers.

However, your drop-outs are not normal. I have TalkTalk FTTC here in Central London last 13 months and apart from Firmware update and TalkTalk Exchange maintenance my connection has not dropped out once. I had problems with disconnections a lot in the past when I was on an ADSL Exchange Only line.

My TalkTalk connection syncs at 70+ days system and connection up time. I did complain in the past that I was experiencing TalkTalk 30 day connection resets. It turned out that my connection wasn't rebooting every 30 days but it was just a superficial reset counter without interrupting my connection so I was actually mistaken.

You need to investigate why your connection is dropping out even if it is once a week. Also at least when it does you should see if the connection is still syncing at 80/20 Mbps in your router stats. If connection still maintains max speeds then at least you can rule out that these drop-outs are not related to noise margin issues.

As for FTTP, if you're in London especially in a block or high rise building, there's a very good chance that you'll get Community Fibre or Hyperoptic soon! My other 2 apartments that I rent out in Central London where I visited to collect letters recently. Got letters from February that both Hyperoptic and Community Fibre have had agreements granted. Openreach FTTP however, is less likely since we've got FTTC already. Openreach don't usually upgrade areas from FTTC to FTTP.

In reply to a post by BLaZiNgSPEED:

"This speed tester is hosted on the same network that broadband connections are terminated on, so this should be the closest speed tester for AAISP customers." It will apply for the rest of the ISPs on FTTC as they are on the same Openreach network.

No that is incorrect. Openreach only connect the home to the exchange, what is nicknamed "last mile".

At the exchange there are a number of "wholesale" operators whom connect thousands of exchanges to the internet provider. AAISP offer a choice from BTwholesale or TalkTalkBusiness, and those networks connect to the ISP's data centre.

Each ISP then chooses how much capacity they pay for from the wholesale operator, and how many sessions they receive at their data centre. This is why people used to talk about contention, how many user sessions per wholesale link

Thinkbroadband really needs a picture of how this works.

Edited by jchamier (Wed 10-Mar-21 08:00:40)

Then why does it show on https://speedtest.aa.net.uk/ ?
They claim

"This speed tester is hosted on the same network that broadband connections are terminated on, so this should be the closest speed tester for AAISP customers."

At the bottom of their speed test? I was left with an impression that this was the reason why with this site I always get more consistent 80Mbps speed test results.

https://librespeed.org/ seems to give me similar speed test results, they seem to be clones of each other, they even look the same.

So basically their claim must be misleading. Or perhaps higher speeds because of "No Flash, No Java, No Websocket"? That's the confusion I have because I was thinking if I am a TalkTalk customer and I am getting close to max speeds on an AAISP speed test then the hosted broadband connection is terminated on the same network. What am I missing here?

In reply to a post by BLaZiNgSPEED:

They claim "This speed tester is hosted on the same network that broadband connections are terminated on, so this should be the closest speed tester for AAISP customers."

It is accurate for AAISP customers. Are you with AAISP as your internet provider? Perhaps they should insert a word in the phrase, ie:
"speed tester is hosted on the same network that AAISP broadband connections"

So basically their claim must be misleading.

Its not misleading, it is accurate but assumes you are a customer. If you are with another ISP then your traffic has to cross out of that ISP, across a neutral interconnect site such as LINX, and then into AAISP network to reach the speed test.

Or perhaps higher speeds because of "No Flash, No Java, No Websocket"? That's the confusion I have because I was thinking if I am a TalkTalk customer and I am getting close to max speeds on an AAISP speed test then the hosted broadband connection is terminated on the same network. What am I missing here?

Nothing. If TalkTalk has their own speed test, it would show the speeds to your ISP, not to the internet.

Edited by jchamier (Wed 10-Mar-21 12:34:22)

Yep, speedof.me just came in with 51.89 for my upload, above the profile speed. The download is higher than I normally see.

As jchamier points out different ISP's terminate at different points.

Not all OpenReach connections terminate at the same point.
OpenReach end of the link ends at the exchange.

That statement is very specific to AAISP customers.
Their tester also exaggerates speeds for their own customers. It's pretty useless.

AAISP use the Librespeed speed test that's why they look the same.
They are both inaccurate.
You cannot get 80Mb throughput on an 80Mb FTTC line.

Edited by j0hn83 (Wed 10-Mar-21 13:51:55)