Educational - Need various connection graphs for 3 days

Hey all,

Firstly apologies if this is not okay to post. I normally use this forum anonymously to browse and I'm currently a grad student about to complete their final project but as part of my practical. I'm measuring the broadband quality of different connection types.

So far I have
- FTTC

Unfortunately, I don't know anyone that has an FTTP, GFAST, ASDL, or Virgin Media connection that can assist me with the other graphs. I was wondering if anyone here has graphs for a 3 day period that would be okay sharing them with myself? Please do strip your IP of course I don't want that. I just need to be able to see the difference in latency between connections. I already suspect FTTP latency to be very minimal. My FTTC connection regularly has 20-30ms ping sadly with frequent packet drops.

My brother has FTTP but because it's a BT router I've heard it's impossible to setup BQM with it. I did try but to no avail, my main requirements are just to get an FTTP graph but GFAST/ASDL/Virgin media is a bonus!

Thanks in advance!

Edited by deleted (Mon 17-Aug-20 15:45:53)

Hi, I don't have a BQM but I think you may be missing something very significant from your project.

Are you aware that latency is far more to do with where you are in the country and where the end point is than it has to do with the connection? The difference in connection latency between the various technologies does exist but compared to the difference in distance it is minor.

For example, if you had a neighbour on FTTP then they may say 1 or 2ms lower latency than your FTTC. However, if say you lived in London and the other person lived in Edinburgh then their latency to London based servers would be a lot higher.

You also need to take account of configuration - a FTTC line with Fastpath will have much lower latency than one that is interleaved.

Then there is also the question of ISP routing. The efficiency of an ISP network in routing to the destination you are looking to reach can be quite significant.

On this basis I believe your methodology will be seriously flawed if you don't understand all of these differences as your statistics will be meaningless without taking account of all of the other factors that can affect latency. I think you may need to rethink your project.

3 days BT FTTP

13th August
14th August
15th August

3 days Virgin DOCSIS (FTTP/RFOG)

7th August
8th August
9th August

Both from my home in Edinburgh area.

As Ian mentions the minimum latency depends on where in the country you are.

On Virgin DOCSIS my minimum latency was 19ms.
On Talktalk FTTC the minimum latency was around 18ms.
On Plusnet FTTC the minimum latency was 17-18ms.
On BT FTTC the minimum latency was always 16ms.
I just moved to BT FTTP and the minimum latency is 14ms.

ISP, technology and location all make a difference.

The huge amounts of yellow on my Virgin BQM are tropical of DOCSIS, as is the higher minimum latency.

Edited by j0hn83 (Mon 17-Aug-20 16:40:22)

Hi John. Whilst I don't have a BQM as I have a BT router I do have BT FTTP - we could do a quick compare of ping times.

I am on the South Coast of England in West Sussex. If I ping thinkbroadband.com then my minimum response time is 7ms (it does vary a bit). I am on BT FTTP 150/30.

Comparing that to your BT FTTP BQM it looks like my proximity to London shaves off around 7ms on the ping. Can you ping thinkbroadband.com to confirm?

This shows that it isn't our technology that is making the difference but out location. If the OP doesn't capture a lot more supporting data about location, ISP and line characterestics (primarily for ADSL/VDSL connections) then the data is going to potentially be badly skewed and unlikely to show anything useful.

If we compared your FTTP speed from Edinburgh with someone in London on FTTC fastpath then I suspect the FTTC fastpath would look way better which on the original proposal would result in a belief that FTTC had lower latency than FTTP by a significant margin.

In reply to a post by j0hn83:

The huge amounts of yellow on my Virgin BQM are tropical of DOCSIS, as is the higher minimum latency.

My live shared BQM (link below) shows the same variabililty, but my location on the Surrey/Hants border helps with the latency a small amount.

13.9ms is the very lowest response I get from pingbox1.thinkbroadband.com (the hostname of the BQM).

Distance is probably the biggest factor in latency, totally agree.
That 7ms difference between our locations sounds very believable.

Every single thing I ping goes from me (Edinburgh) to my ISP BT (London) and back again.
That's roughly 600-700km away.

600km x 2 (there and back) = 1200km
Light travels at around 200km/sec in fibre.
1200km ÷ 200000km/s = 0.006s aka 6ms

So latency increases roughly 1ms per 100km.
That's at least 6-7ms latency added to my line from my geography alone.

No idea what route down the country the BT backhaul takes but I'm about 9ms higher latency than those in central London on BT FTTP so either a longer snaking fibre route than I'm guessing or hardware in between adding delay.

Every piece of hardware at every hop between 2 end points (routers, L2S, OLT's, etc) can add additional delay.

In my experience and what I've observed from others is FTTC has roughly 2-3ms higher latency compared to FTTP, if from the same location, on the same ISP, with the same routing.

xDSL technologies (VDSL and ADSL) can also add deliberate additional delay in the form of Interleaving.
That can make the difference in latency between FTTC and FTTP much higher.

edit: most of this is for the OP's benefit. Not trying to lecture anyone

Edited by j0hn83 (Mon 17-Aug-20 18:24:07)

In reply to a post by ian72:

Hi, I don't have a BQM but I think you may be missing something very significant from your project.

Are you aware that latency is far more to do with where you are in the country and where the end point is than it has to do with the connection? The difference in connection latency between the various technologies does exist but compared to the difference in distance it is minor.

For example, if you had a neighbour on FTTP then they may say 1 or 2ms lower latency than your FTTC. However, if say you lived in London and the other person lived in Edinburgh then their latency to London based servers would be a lot higher.

You also need to take account of configuration - a FTTC line with Fastpath will have much lower latency than one that is interleaved.

Then there is also the question of ISP routing. The efficiency of an ISP network in routing to the destination you are looking to reach can be quite significant.

On this basis I believe your methodology will be seriously flawed if you don't understand all of these differences as your statistics will be meaningless without taking account of all of the other factors that can affect latency. I think you may need to rethink your project.

The first one I was, of course, aware of it's the same reason why EU gamers have horrid ping on US-based gaming servers and vice versa.

I wasn't aware of FTTC line with fastpath - I'll be sure to investigate that because that sounds rather interesting.

The point of ISP routing is a good point but I have mentioned that in my project. These graphs and my investigation of them is noted with more of a "grain of salt" type of approach and clearly disclosing those relevant key points. I am stating as much information where possible but I totally understand your concerns and thanks for the help!

In reply to a post by j0hn83:

3 days BT FTTP

13th August
14th August
15th August

3 days Virgin DOCSIS (FTTP/RFOG)

7th August
8th August
9th August

Both from my home in Edinburgh area.

As Ian mentions the minimum latency depends on where in the country you are.

On Virgin DOCSIS my minimum latency was 19ms.
On Talktalk FTTC the minimum latency was around 18ms.
On Plusnet FTTC the minimum latency was 17-18ms.
On BT FTTC the minimum latency was always 16ms.
I just moved to BT FTTP and the minimum latency is 14ms.

ISP, technology and location all make a difference.

The huge amounts of yellow on my Virgin BQM are tropical of DOCSIS, as is the higher minimum latency.

Great and that's even extra information to help support my project - providing you would be okay wih me including about the latency information?

In reply to a post by ian72:

Hi John. Whilst I don't have a BQM as I have a BT router I do have BT FTTP - we could do a quick compare of ping times.

I am on the South Coast of England in West Sussex. If I ping thinkbroadband.com then my minimum response time is 7ms (it does vary a bit). I am on BT FTTP 150/30.

Comparing that to your BT FTTP BQM it looks like my proximity to London shaves off around 7ms on the ping. Can you ping thinkbroadband.com to confirm?

This shows that it isn't our technology that is making the difference but out location. If the OP doesn't capture a lot more supporting data about location, ISP and line characterestics (primarily for ADSL/VDSL connections) then the data is going to potentially be badly skewed and unlikely to show anything useful.

If we compared your FTTP speed from Edinburgh with someone in London on FTTC fastpath then I suspect the FTTC fastpath would look way better which on the original proposal would result in a belief that FTTC had lower latency than FTTP by a significant margin.

That would be hugely appreciated!

EDIT: I wanted to thank everyone for their responses so far, you have all been a huge help I may not understand extreme technicals I'd say I'm about a rookie status but I will still be sure to try my best. Of course the more data I have to sift through the stronger my report gets so I have no objection to any/all extra data

Edited by deleted (Mon 17-Aug-20 18:35:24)

In reply to a post by Peter5000:

The first one I was, of course, aware of it's the same reason why EU gamers have horrid ping on US-based gaming servers and vice versa.

I have never had any issues playing on US server when playing games with my US friends, I get about 98ms to the US East Coast and about 149ms to some servers located in the west coast.
In fact I have a few times had better ping / latency on a US Server than a few of my US friends, I was about 5 to 10ms lower than them, but the US is a very large place.

I also think I am not that far from where the TBB Servers are being hosted due to I get a very low ping to them shown below which also shows in my BQM in my sig:

ping pingbox1.thinkbroadband.com

Pinging pingbox1.thinkbroadband.com [80.249.99.164] with 32 bytes of data:
Reply from 80.249.99.164: bytes=32 time=2ms TTL=56
Reply from 80.249.99.164: bytes=32 time=2ms TTL=56
Reply from 80.249.99.164: bytes=32 time=3ms TTL=56
Reply from 80.249.99.164: bytes=32 time=3ms TTL=56

Ping statistics for 80.249.99.164:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 2ms, Maximum = 3ms, Average = 2ms

Paul

thinkbroadband, hosted by Netconnex, is I believe sited in a couple of the Telehouse peering datacentres in Docklands.

Edited by RobertoS (Tue 18-Aug-20 01:56:05)

In reply to a post by PaulKirby:

I also think I am not that far from where the TBB Servers are being hosted due to I get a very low ping to them shown below which also shows in my BQM in my sig:

Your .sig says you're in Ilford, which I think is London, as are the TBB servers. So you have no physical distance to cover to the TBB servers, just the variabillity of how the fibres and networks are inter-connected.