Discussion:
ADSL2+ line sync data
(too old to reply)
Paul Brooks
2013-09-12 02:17:58 UTC
Permalink
A recent Ofcom (UK) report has a very interesting chart of ADSL2+ line speeds:

Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
Figure 4 on page 11:

This chart is effectively the result of the ADSL line-sync/attenuation curve combined
with the increasing area of circles of increasing radius around the exchange - and
demonstrates very clearly why so many people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)

Also scatter-plots of sync-speed with line-length, as per Figure 8 from another UK report:


Now every DSLAM network operator can put together similar charts - but I'm not aware
of any stats for Australian networks, apart from the heat maps put out by iiNet and
the adsl2exchanges.com.au site, which aren't quite what I'm looking for.

For all you DSLAM operators - I would be very interested in putting together similar
charts for the Australian networks, to see how our copper loop network varies from the
UK network. If anyone is willing to share data or statistics, I'm very interested in
pulling together similar Australian charts - on a non-identified, aggregated,
anonymised basis if you wish.
Please contact me off-list - thanks.

Paul.
Tim March
2013-09-12 02:27:41 UTC
Permalink
Slightly OT, but I had a bunch of DSL issues at my home recently and
fired up Cacti on my Synology NAS (I think I mentioned this in a post a
few weeks back...) to graph DSL statistics on a Cisco 867W.

Relevant links if anyone is interested in doing the same thing;

http://www.robdehoog.nl/technology/cacti-on-a-synology-nas-how-to/

http://www.alcatron.net/?p=42

http://itknowledgeexchange.techtarget.com/cisco/understanding-cisco-dsl-statistics/



T.
Post by Paul Brooks
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
This chart is effectively the result of the ADSL line-sync/attenuation curve
combined with the increasing area of circles of increasing radius around the
exchange - and demonstrates very clearly why so many people get low ADSL2+ line
speeds.
(Exercise for the reader - work out how VDSL2 would be any different)
Now every DSLAM network operator can put together similar charts - but I'm not
aware of any stats for Australian networks, apart from the heat maps put out by
iiNet and the adsl2exchanges.com.au site, which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting together
similar charts for the Australian networks, to see how our copper loop network
varies from the UK network. If anyone is willing to share data or statistics,
I'm very interested in pulling together similar Australian charts - on a
non-identified, aggregated, anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
--
PGP/GNUPG Public Key: http://d3vnu11.com/pub.key
Geordie Guy
2013-09-12 03:08:52 UTC
Permalink
That first link is blocked by our Barracuda filter due to spyware?
Post by Tim March
Slightly OT, but I had a bunch of DSL issues at my home recently and
fired up Cacti on my Synology NAS (I think I mentioned this in a post a
few weeks back...) to graph DSL statistics on a Cisco 867W.
Relevant links if anyone is interested in doing the same thing;
http://www.robdehoog.nl/technology/cacti-on-a-synology-nas-how-to/
http://www.alcatron.net/?p=42
http://itknowledgeexchange.techtarget.com/cisco/understanding-cisco-dsl-statistics/
T.
Post by Paul Brooks
A recent Ofcom (UK) report has a very interesting chart of ADSL2+ line
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
Post by Paul Brooks
This chart is effectively the result of the ADSL line-sync/attenuation
curve
Post by Paul Brooks
combined with the increasing area of circles of increasing radius around
the
Post by Paul Brooks
exchange - and demonstrates very clearly why so many people get low
ADSL2+ line
Post by Paul Brooks
speeds.
(Exercise for the reader - work out how VDSL2 would be any different)
Also scatter-plots of sync-speed with line-length, as per Figure 8 from
another
Post by Paul Brooks
Now every DSLAM network operator can put together similar charts - but
I'm not
Post by Paul Brooks
aware of any stats for Australian networks, apart from the heat maps put
out by
Post by Paul Brooks
iiNet and the adsl2exchanges.com.au site, which aren't quite what I'm
looking for.
Post by Paul Brooks
For all you DSLAM operators - I would be very interested in putting
together
Post by Paul Brooks
similar charts for the Australian networks, to see how our copper loop
network
Post by Paul Brooks
varies from the UK network. If anyone is willing to share data or
statistics,
Post by Paul Brooks
I'm very interested in pulling together similar Australian charts - on a
non-identified, aggregated, anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
--
PGP/GNUPG Public Key: http://d3vnu11.com/pub.key
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Matt Perkins
2013-09-12 02:51:41 UTC
Permalink
I think just taking distance into account is very simplistic view of
things. Im not sure what state the copper is in the UK. However in
Australia there are other things to consider other then distance.

Joints. - The number of joints in an adsl loop dramatically decrease SNR
and increase loss. The type of joint is also of interest. Insulation
displacement is far better then non soldered pin strip for example.

Cable type / gauge. - Cable type and gauge varies greatly. There is far
less loss on higher gauge cable. However most of the time gauge changes
at every joint this produces all sorts of nasty results from impedance
mismatches to standing waves.

Bridge Tap - xDSL's worst enemy and Telecom stalwart from the 1970-80's
bridged taps are responsible for most of the SNR problems on xDSL. I
remember pulling these things out for a living when attempting to reduce
the BERT on Auspac and X25 services back in the 80's (now im showing my
age) These things are going to be the biggest problem for the new
VDSL/FTTN environment. Technicians paid per completed service will not
have a lot of incentive to go find and remove these so the rule of thumb
will be. "Does it work at any speed, Yes. Good next job" . Just as it is
now.

Customer premiss cabling - Although this isn't some time's our problem
as ISP's it becomes our problem through support man hours and unhappy
customers, How many customers have you had a discussion with about that
old K-Mart non twisted pair 20m extension cord and those old disused 5
parallel sockets (more bridged tap's). It cant be my extension cord it
worked yesterday.

<rant>
My experience is the people working in this space, maintaining the
Telstra CAN (or customer area network) have very poor understanding of
these issues and the ones that do dont have the time to care when they
apply the pizza delivery pay model to these guys. None of this will
change, price pressure and time frame's will make a network that needs
serious TLC even worse. I have a customer that has been waiting 7 months
for copper that was damaged during a road works project to be replaced.
3G for 7 months. Good luck with that.

VDSL has it's place. In buildings - however we spend serious time going
to customers sites in buildings only to find that a technician with a
time pressure was there that morning and stole what they thought was a
spare pair for a phone service. Just because they could hear no dial
tone on it. No dial tone it must be free.

Welcome to the brave new - old world of data over copper.
</rant>
Post by Paul Brooks
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
This chart is effectively the result of the ADSL line-sync/attenuation
curve combined with the increasing area of circles of increasing
radius around the exchange - and demonstrates very clearly why so many
people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)
Now every DSLAM network operator can put together similar charts - but
I'm not aware of any stats for Australian networks, apart from the
heat maps put out by iiNet and the adsl2exchanges.com.au site, which
aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting
together similar charts for the Australian networks, to see how our
copper loop network varies from the UK network. If anyone is willing
to share data or statistics, I'm very interested in pulling together
similar Australian charts - on a non-identified, aggregated,
anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Paul Wallace
2013-09-12 04:44:30 UTC
Permalink
The dulcet tones of wisdom come from maturity and usually maturity alone Matt!






On 12/09/2013, at 12:52 PM, "Matt Perkins" <***@spectrum.com.au<mailto:***@spectrum.com.au>> wrote:

I think just taking distance into account is very simplistic view of things. Im not sure what state the copper is in the UK. However in Australia there are other things to consider other then distance.

Joints. - The number of joints in an adsl loop dramatically decrease SNR and increase loss. The type of joint is also of interest. Insulation displacement is far better then non soldered pin strip for example.

Cable type / gauge. - Cable type and gauge varies greatly. There is far less loss on higher gauge cable. However most of the time gauge changes at every joint this produces all sorts of nasty results from impedance mismatches to standing waves.

Bridge Tap - xDSL's worst enemy and Telecom stalwart from the 1970-80's bridged taps are responsible for most of the SNR problems on xDSL. I remember pulling these things out for a living when attempting to reduce the BERT on Auspac and X25 services back in the 80's (now im showing my age) These things are going to be the biggest problem for the new VDSL/FTTN environment. Technicians paid per completed service will not have a lot of incentive to go find and remove these so the rule of thumb will be. "Does it work at any speed, Yes. Good next job" . Just as it is now.

Customer premiss cabling - Although this isn't some time's our problem as ISP's it becomes our problem through support man hours and unhappy customers, How many customers have you had a discussion with about that old K-Mart non twisted pair 20m extension cord and those old disused 5 parallel sockets (more bridged tap's). It cant be my extension cord it worked yesterday.

<rant>
My experience is the people working in this space, maintaining the Telstra CAN (or customer area network) have very poor understanding of these issues and the ones that do dont have the time to care when they apply the pizza delivery pay model to these guys. None of this will change, price pressure and time frame's will make a network that needs serious TLC even worse. I have a customer that has been waiting 7 months for copper that was damaged during a road works project to be replaced. 3G for 7 months. Good luck with that.

VDSL has it's place. In buildings - however we spend serious time going to customers sites in buildings only to find that a technician with a time pressure was there that morning and stole what they thought was a spare pair for a phone service. Just because they could hear no dial tone on it. No dial tone it must be free.

Welcome to the brave new - old world of data over copper.
</rant>



On 12/09/13 12:17 PM, Paul Brooks wrote:
A recent Ofcom (UK) report has a very interesting chart of ADSL2+ line speeds:
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
Figure 4 on page 11:
<ATT00001.png>
This chart is effectively the result of the ADSL line-sync/attenuation curve combined with the increasing area of circles of increasing radius around the exchange - and demonstrates very clearly why so many people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)

Also scatter-plots of sync-speed with line-length, as per Figure 8 from another UK report:

<ATT00002.png>
Now every DSLAM network operator can put together similar charts - but I'm not aware of any stats for Australian networks, apart from the heat maps put out by iiNet and the adsl2exchanges.com.au<http://adsl2exchanges.com.au> site, which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting together similar charts for the Australian networks, to see how our copper loop network varies from the UK network. If anyone is willing to share data or statistics, I'm very interested in pulling together similar Australian charts - on a non-identified, aggregated, anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.






_______________________________________________
AusNOG mailing list
***@lists.ausnog.net<mailto:***@lists.ausnog.net>
http://lists.ausnog.net/mailman/listinfo/ausnog


_______________________________________________
AusNOG mailing list
***@lists.ausnog.net<mailto:***@lists.ausnog.net>
http://lists.ausnog.net/mailman/listinfo/ausnog
Beeson, Ayden
2013-09-12 03:45:39 UTC
Permalink
I've found the attenuation to be the best stat for general information, all the below items contribute to it, compare it to distance and you'll get a good measure of copper quality etc.

Matt Perkins <***@spectrum.com.au> wrote:



I think just taking distance into account is very simplistic view of things. Im not sure what state the copper is in the UK. However in Australia there are other things to consider other then distance.

Joints. - The number of joints in an adsl loop dramatically decrease SNR and increase loss. The type of joint is also of interest. Insulation displacement is far better then non soldered pin strip for example.

Cable type / gauge. - Cable type and gauge varies greatly. There is far less loss on higher gauge cable. However most of the time gauge changes at every joint this produces all sorts of nasty results from impedance mismatches to standing waves.

Bridge Tap - xDSL's worst enemy and Telecom stalwart from the 1970-80's bridged taps are responsible for most of the SNR problems on xDSL. I remember pulling these things out for a living when attempting to reduce the BERT on Auspac and X25 services back in the 80's (now im showing my age) These things are going to be the biggest problem for the new VDSL/FTTN environment. Technicians paid per completed service will not have a lot of incentive to go find and remove these so the rule of thumb will be. "Does it work at any speed, Yes. Good next job" . Just as it is now.

Customer premiss cabling - Although this isn't some time's our problem as ISP's it becomes our problem through support man hours and unhappy customers, How many customers have you had a discussion with about that old K-Mart non twisted pair 20m extension cord and those old disused 5 parallel sockets (more bridged tap's). It cant be my extension cord it worked yesterday.

<rant>
My experience is the people working in this space, maintaining the Telstra CAN (or customer area network) have very poor understanding of these issues and the ones that do dont have the time to care when they apply the pizza delivery pay model to these guys. None of this will change, price pressure and time frame's will make a network that needs serious TLC even worse. I have a customer that has been waiting 7 months for copper that was damaged during a road works project to be replaced. 3G for 7 months. Good luck with that.

VDSL has it's place. In buildings - however we spend serious time going to customers sites in buildings only to find that a technician with a time pressure was there that morning and stole what they thought was a spare pair for a phone service. Just because they could hear no dial tone on it. No dial tone it must be free.

Welcome to the brave new - old world of data over copper.
</rant>



On 12/09/13 12:17 PM, Paul Brooks wrote:
A recent Ofcom (UK) report has a very interesting chart of ADSL2+ line speeds:
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
Figure 4 on page 11:
[cid:***@spectrum.com.au]
This chart is effectively the result of the ADSL line-sync/attenuation curve combined with the increasing area of circles of increasing radius around the exchange - and demonstrates very clearly why so many people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)

Also scatter-plots of sync-speed with line-length, as per Figure 8 from another UK report:

[cid:***@spectrum.com.au]
Now every DSLAM network operator can put together similar charts - but I'm not aware of any stats for Australian networks, apart from the heat maps put out by iiNet and the adsl2exchanges.com.au site, which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting together similar charts for the Australian networks, to see how our copper loop network varies from the UK network. If anyone is willing to share data or statistics, I'm very interested in pulling together similar Australian charts - on a non-identified, aggregated, anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.






_______________________________________________
AusNOG mailing list
***@lists.ausnog.net<mailto:***@lists.ausnog.net>
http://lists.ausnog.net/mailman/listinfo/ausnog



[cid:csu-logo3699.bmp]<http://www.csu.edu.au/>

| ALBURY-WODONGA | BATHURST | CANBERRA | DUBBO | GOULBURN | MELBOURNE | ONTARIO | ORANGE | PORT MACQUARIE | SYDNEY | WAGGA WAGGA |

________________________________
LEGAL NOTICE
This email (and any attachment) is confidential and is intended for the use of the addressee(s) only. If you are not the intended recipient of this email, you must not copy, distribute, take any action in reliance on it or disclose it to anyone. Any confidentiality is not waived or lost by reason of mistaken delivery. Email should be checked for viruses and defects before opening. Charles Sturt University (CSU) does not accept liability for viruses or any consequence which arise as a result of this email transmission. Email communications with CSU may be subject to automated email filtering, which could result in the delay or deletion of a legitimate email before it is read at CSU. The views expressed in this email are not necessarily those of CSU.

Charles Sturt University in Australia<http://www.csu.edu.au> The Grange Chancellery, Panorama Avenue, Bathurst NSW Australia 2795 (ABN: 83 878 708 551; CRICOS Provider Numbers: 00005F (NSW), 01947G (VIC), 02960B (ACT)). TEQSA Provider Number: PV12018
Charles Sturt University in Ontario<http://www.charlessturt.ca/> 860 Harrington Court, Burlington Ontario Canada L7N 3N4 Registration: www.peqab.ca<http://www.peqab.ca>

Consider the environment before printing this email.

Disclaimer added by CodeTwo Exchange Rules 2007
www.codetwo.com<http://www.codetwo.com>
Guy Ellis
2013-09-12 21:08:38 UTC
Permalink
Paul,

In response to your challenge (Exercise for the reader - work out how
VDSL2 would be any different)...

In contrast to the current ADSL2+ network, there are 3 big differences
with the proposed VDSL2 FTTN deployment -
(i) shorter loop lengths (700-800m)
(ii) vectoring (crosstalk--)
(iii) bonding (speed++)

While such a VDSL2 network is not as good as fibre, it's no where near
as bad as the current ADSL2+ network.
Right now some poor folks are on 6km loop lengths, there's plenty of
crosstalk and getting bonding working is a challenge.

Regards,
- Guy.
Post by Paul Brooks
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
This chart is effectively the result of the ADSL line-sync/attenuation
curve combined with the increasing area of circles of increasing
radius around the exchange - and demonstrates very clearly why so many
people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)
Now every DSLAM network operator can put together similar charts - but
I'm not aware of any stats for Australian networks, apart from the
heat maps put out by iiNet and the adsl2exchanges.com.au site, which
aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting
together similar charts for the Australian networks, to see how our
copper loop network varies from the UK network. If anyone is willing
to share data or statistics, I'm very interested in pulling together
similar Australian charts - on a non-identified, aggregated,
anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
--
Guy Ellis
***@traverse.com.au
www.traverse.com.au
T: +61 3 9386 4435 M: +61 419 398 234
Paul Brooks
2013-09-12 23:02:36 UTC
Permalink
Post by Beeson, Ayden
Paul,
In response to your challenge (Exercise for the reader - work out how VDSL2 would be
any different)...
In contrast to the current ADSL2+ network, there are 3 big differences with the
proposed VDSL2 FTTN deployment -
(i) shorter loop lengths (700-800m)
(ii) vectoring (crosstalk--)
(iii) bonding (speed++)
While such a VDSL2 network is not as good as fibre, it's no where near as bad as the
current ADSL2+ network.
Right now some poor folks are on 6km loop lengths, there's plenty of crosstalk and
getting bonding working is a challenge.
You can forget about pair bonding in the broader plan - I sincerely doubt the budget
or street cabinets will be big enough to build two ports for every dwelling.

My point was that - without vectoring - the VDSL2 chart would look much the same.

Sure the axis labels would change - distances to 5km become distances to 800 metres,
bandwidth tops out at 120 Mbps instead of 24 Mbps - but the shape of the chart would
look much the same. A negligable proportion getting the full 'up to' speed, roughly
30% of people getting speeds down to 50% of the "up to" limit, and a large hump
majority of people down the low end getting about 10 - 15 Mbps - probably better than
the ~4 Mbps they might get now with ADSL2, but not really up to the new benchmark.

That leaves vectoring as the major difference - which will make speeds more
predictable and push a lot more services to the right to higher speeds, reduce the
width of the fuzzy cloud in the second diagram, but still won't deliver 50 Mbps
further than about 750 metres.


Anyway, back to the original topic - I was looking for ADSL2+ data - anyone?
Post by Beeson, Ayden
Regards,
- Guy.
Post by Paul Brooks
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
This chart is effectively the result of the ADSL line-sync/attenuation curve
combined with the increasing area of circles of increasing radius around the
exchange - and demonstrates very clearly why so many people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)
Also scatter-plots of sync-speed with line-length, as per Figure 8 from another UK
Now every DSLAM network operator can put together similar charts - but I'm not
aware of any stats for Australian networks, apart from the heat maps put out by
iiNet and the adsl2exchanges.com.au site, which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting together
similar charts for the Australian networks, to see how our copper loop network
varies from the UK network. If anyone is willing to share data or statistics, I'm
very interested in pulling together similar Australian charts - on a
non-identified, aggregated, anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
--
Guy Ellis
www.traverse.com.au
T: +61 3 9386 4435 M: +61 419 398 234
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Mike Trewartha
2013-09-12 23:16:06 UTC
Permalink
With the change in loop lengths, what is the likelihood of some bodies current decent (ie. 18mbit+) ADSL2+ sync speeds dropping once FTTN is deployed?


Regards, Mike.

Sent from my iPhone
Post by Beeson, Ayden
Paul,
In response to your challenge (Exercise for the reader - work out how VDSL2 would be any different)...
In contrast to the current ADSL2+ network, there are 3 big differences with the proposed VDSL2 FTTN deployment -
(i) shorter loop lengths (700-800m)
(ii) vectoring (crosstalk--)
(iii) bonding (speed++)
While such a VDSL2 network is not as good as fibre, it's no where near as bad as the current ADSL2+ network.
Right now some poor folks are on 6km loop lengths, there's plenty of crosstalk and getting bonding working is a challenge.
You can forget about pair bonding in the broader plan - I sincerely doubt the budget or street cabinets will be big enough to build two ports for every dwelling.
My point was that - without vectoring - the VDSL2 chart would look much the same.
Sure the axis labels would change - distances to 5km become distances to 800 metres, bandwidth tops out at 120 Mbps instead of 24 Mbps - but the shape of the chart would look much the same. A negligable proportion getting the full 'up to' speed, roughly 30% of people getting speeds down to 50% of the "up to" limit, and a large hump majority of people down the low end getting about 10 - 15 Mbps - probably better than the ~4 Mbps they might get now with ADSL2, but not really up to the new benchmark.
That leaves vectoring as the major difference - which will make speeds more predictable and push a lot more services to the right to higher speeds, reduce the width of the fuzzy cloud in the second diagram, but still won't deliver 50 Mbps further than about 750 metres.
Anyway, back to the original topic - I was looking for ADSL2+ data - anyone?
Post by Beeson, Ayden
Regards,
- Guy.
Post by Paul Brooks
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
<mime-attachment.png>
This chart is effectively the result of the ADSL line-sync/attenuation curve combined with the increasing area of circles of increasing radius around the exchange - and demonstrates very clearly why so many people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)
<mime-attachment.png>
Now every DSLAM network operator can put together similar charts - but I'm not aware of any stats for Australian networks, apart from the heat maps put out by iiNet and the adsl2exchanges.com.au site, which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting together similar charts for the Australian networks, to see how our copper loop network varies from the UK network. If anyone is willing to share data or statistics, I'm very interested in pulling together similar Australian charts - on a non-identified, aggregated, anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
--
Guy Ellis
www.traverse.com.au
T: +61 3 9386 4435 M: +61 419 398 234
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Tony
2013-09-12 23:27:48 UTC
Permalink
I would have thought minimal chance. If they're that close to an exchange/RIM to get good sync/speed, then they should remain that close with FTTN. It's possible there might be some silliness that happens due to cable paths, but it would have to be a fairly corner case where speed would drop. The copper isn't being changed, just what it connects to.


regards,
Tony.
Post by Paul Brooks
________________________________
Sent: Friday, 13 September 2013 9:16 AM
Subject: Re: [AusNOG] ADSL2+ line sync data
With the change in loop lengths, what is the likelihood of some bodies current decent (ie. 18mbit+) ADSL2+ sync speeds dropping once FTTN is deployed?
Regards, Mike. 
Sent from my iPhone
Post by Paul Brooks
Paul,
Post by Guy Ellis
In response to your challenge (Exercise for the reader - work
out how VDSL2 would be any different)...
Post by Paul Brooks
Post by Paul Brooks
Post by Guy Ellis
In contrast to the current ADSL2+ network, there are 3 big
differences with the proposed VDSL2 FTTN deployment -
Post by Paul Brooks
Post by Paul Brooks
Post by Guy Ellis
(i) shorter loop lengths (700-800m)
(ii) vectoring (crosstalk--)
(iii) bonding (speed++)
While such a VDSL2 network is not as good as fibre, it's no
where near as bad as the current ADSL2+ network.
Post by Paul Brooks
Post by Paul Brooks
Post by Guy Ellis
Right now some poor folks are on 6km loop lengths, there's
plenty of crosstalk and getting bonding working is a challenge.
You can forget about pair bonding in the broader plan - I sincerely doubt the budget or street cabinets will be big enough to build two ports for every dwelling.
Post by Paul Brooks
Post by Paul Brooks
My point was that - without vectoring - the VDSL2 chart would look much the same.
Sure the axis labels would change - distances to 5km become
distances to 800 metres, bandwidth tops out at 120 Mbps instead of
24 Mbps - but the shape of the chart would look much the same. A
negligable proportion getting the full 'up to' speed, roughly 30% of
people getting speeds down to 50% of the "up to" limit, and a large
hump majority of people down the low end getting about 10 - 15 Mbps
- probably better than the ~4 Mbps they might get now with ADSL2,
but not really up to the new benchmark.
Post by Paul Brooks
Post by Paul Brooks
That leaves vectoring as the major difference - which will make
speeds more predictable and push a lot more services to the right to
higher speeds, reduce the width of the fuzzy cloud in the second
diagram, but still won't deliver 50 Mbps further than about 750
metres.
Post by Paul Brooks
Post by Paul Brooks
Anyway, back to the original topic  - I was looking for ADSL2+ data
- anyone?
Post by Paul Brooks
Post by Paul Brooks
 
Post by Guy Ellis
Regards,
 - Guy.
Post by Paul Brooks
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
<mime-attachment.png>
This chart is effectively the result of the ADSL line-sync/attenuation curve combined with the increasing area of circles of increasing radius around the exchange - and demonstrates very clearly why so many people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be
any different)
Post by Paul Brooks
Post by Paul Brooks
Post by Guy Ellis
Post by Paul Brooks
<mime-attachment.png>
Now every DSLAM network operator can put together similar charts - but I'm not aware of any stats for Australian networks, apart from the heat maps put out by iiNet and the adsl2exchanges.com.au site, which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting together similar charts for the Australian networks, to see how our copper loop network varies from the UK network. If anyone is willing to share data or statistics, I'm very interested in pulling together similar Australian charts - on a non-identified, aggregated, anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.
 
_______________________________________________
--
_______________________________________________
_______________________________________________
Post by Paul Brooks
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Paul Brooks
2013-09-12 23:53:59 UTC
Permalink
Post by Mike Trewartha
With the change in loop lengths, what is the likelihood of some bodies current
decent (ie. 18mbit+) ADSL2+ sync speeds dropping once FTTN is deployed?
None at all I expect. Loop lengths won't get any longer. If they are getting 18 Mbps
on ADSL2+, they'll get higher with VDSL2 even if their loop length doesn't change.

P.
Peter Adkins
2013-09-13 00:40:33 UTC
Permalink
There's a couple of papers available on the IEEE around FEXT with regards
to ADSL2+ services. More specifically, how much of an impact cross-talk can
have within an environment where a large number of surrounding pairs in a
bundle are also used to provide an ADSL2+ service.

http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4446136
Matt Perkins
2013-09-13 01:33:54 UTC
Permalink
I think there are to many unknown's to apply much of this type of logic
to NBN2. The 100M/bit question will be.

Where will they put the "node". The most logical place will be to knock
over the pillars (big grey knob looking post) and put the big green
graffiti target in it's place. The Node will then cover what was
formally known as the "distribution area" (DA).

If that's the case the direct main cable usually 300+ pairs direct to
the exchange will be superseded with the fiber & node. This higher
density cable is where most of the cross talk is seen usually. It's
induced in these cables but the root cause is usually impedance
mismatch further down the loop. With the higher density cable out of the
picture I think the cross talk will be decreased overall but the root
cause will still be there to cause havoc.

Warning detail ahead !

Iam assuming BYO CPE and internal customer premisses cable . So the
mismatches will still be there even if someone had the foresight to go
and remove all the bridged tap's in the joining posts and pits. Most of
these exist in large epoxy filled heads that is basically just 3 pairs
scotch locked together stuck in what looks like the bottom half of a 2l
drink bottle and filled with Epoxy. Some even have passive L/C
(inductors/coils of wire) components in place that were used to balance
the loop back toward the now non existing mains cable. Records for this
stuff is flakey to non-existent. It has to be found. Burred in the
ground in unknown locations encapsulated in epoxy some in asbestos pits.
Some of these pits have been covered by earth or concrete years ago.
There is little financial case to make all this happen. It's man power
intensive. Time consuming dirty work. It does not fit the Pizza delivery
business model and wont happen. You may as well put in fiber at that
point.

If i was tasked with this brief. That is to get this thing working as
quickly and cheaply as possible so it can be sold. This is what I would do.

Get a contractor to build a custom node housing. It would be a big box
that would allow the dslam/vdsl equipment at one end and the other end
would fit over the existing pillar. Jumpers could then be run from the
dslam to the pillar untill all subscribers were off the main's cable.
The mains cable could then be ripped out by a disposal contractor to
sell the copper.

Nothing would be done further down the loop. It would work how it works
now best effort. Customer would go to dick smith and get themselves a
vdsl modem with POTS/VOIP port and that would be that. Average speed
would be <25M/bit with a the lucky few <50M/bit

It would be very quick to finish. Does not touch the ends of the CAN
where all the problems are and the whole thing could be sold to a sucker
(aka mum&dad's via share offers and super funds) and we can go through
this whole thing again in 10 years.

Matt.
Post by Peter Adkins
There's a couple of papers available on the IEEE around FEXT with
regards to ADSL2+ services. More specifically, how much of an impact
cross-talk can have within an environment where a large number of
surrounding pairs in a bundle are also used to provide an ADSL2+ service.
http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4446136
Paul Brooks
2013-09-13 02:10:04 UTC
Permalink
I think there are to many unknown's to apply much of this type of logic to NBN2. The
100M/bit question will be.
Where will they put the "node". The most logical place will be to knock over the
pillars (big grey knob looking post) and put the big green graffiti target in it's
place. The Node will then cover what was formally known as the "distribution area" (DA).
Knocking it over is likely to result in too long without any service for many hundreds
of people. Most proposals I've seen assume the pillar itself stays, the node cabinet
is installed beside it, and 100-pair cables from the MDF in the cabinet are run into
the pillar, replacing the main cables from the exchange. The customer pairs are
cross-connected within the pillar as normal.
That's what they did in the UK, and the new regime seems to like to follow the UK
model in many things.
If that's the case the direct main cable usually 300+ pairs direct to the exchange
will be superseded with the fiber & node.
Yes.
This higher density cable is where most of the cross talk is seen usually. It's
induced in these cables but the root cause is usually impedance mismatch further
down the loop. With the higher density cable out of the picture I think the cross
talk will be decreased overall but the root cause will still be there to cause havoc.
Vectoring fixes this.
Nothing would be done further down the loop. It would work how it works now best
effort. Customer would go to dick smith and get themselves a vdsl modem with
POTS/VOIP port and that would be that. Average speed would be <25M/bit with a the
lucky few <50M/bit
except the election promise was 25 Mbps minimum for all to begin with, raising to 50
Mbps minimum for 90% of people after vectoring is installed.
I can see a lot of truckrolls to homes to install VDSL2-capable central
filters/splitters - BT in the UK also confirms that many of the problems woith VDSL2
performance they've seen relate to in-home cabling.
Its still going to be a good time to be a registered cabling contractor.

P.
Guy Ellis
2013-09-12 23:37:50 UTC
Permalink
According to Turnbull, bonding will be supported, but I'll conceed that
this is not a residential service.
I think most a business premises have at least two copper pairs?

Vectoring will need to become mandatory at some point to reach the
50Mbps target of the policy.

- G.
Post by Paul Brooks
Post by Guy Ellis
Paul,
In response to your challenge (Exercise for the reader - work out how
VDSL2 would be any different)...
In contrast to the current ADSL2+ network, there are 3 big
differences with the proposed VDSL2 FTTN deployment -
(i) shorter loop lengths (700-800m)
(ii) vectoring (crosstalk--)
(iii) bonding (speed++)
While such a VDSL2 network is not as good as fibre, it's no where
near as bad as the current ADSL2+ network.
Right now some poor folks are on 6km loop lengths, there's plenty of
crosstalk and getting bonding working is a challenge.
You can forget about pair bonding in the broader plan - I sincerely
doubt the budget or street cabinets will be big enough to build two
ports for every dwelling.
My point was that - without vectoring - the VDSL2 chart would look much the same.
Sure the axis labels would change - distances to 5km become distances
to 800 metres, bandwidth tops out at 120 Mbps instead of 24 Mbps - but
the shape of the chart would look much the same. A negligable
proportion getting the full 'up to' speed, roughly 30% of people
getting speeds down to 50% of the "up to" limit, and a large hump
majority of people down the low end getting about 10 - 15 Mbps -
probably better than the ~4 Mbps they might get now with ADSL2, but
not really up to the new benchmark.
That leaves vectoring as the major difference - which will make speeds
more predictable and push a lot more services to the right to higher
speeds, reduce the width of the fuzzy cloud in the second diagram, but
still won't deliver 50 Mbps further than about 750 metres.
Anyway, back to the original topic - I was looking for ADSL2+ data - anyone?
Post by Guy Ellis
Regards,
- Guy.
Post by Paul Brooks
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
This chart is effectively the result of the ADSL
line-sync/attenuation curve combined with the increasing area of
circles of increasing radius around the exchange - and demonstrates
very clearly why so many people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)
Also scatter-plots of sync-speed with line-length, as per Figure 8
Now every DSLAM network operator can put together similar charts -
but I'm not aware of any stats for Australian networks, apart from
the heat maps put out by iiNet and the adsl2exchanges.com.au site,
which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting
together similar charts for the Australian networks, to see how our
copper loop network varies from the UK network. If anyone is willing
to share data or statistics, I'm very interested in pulling together
similar Australian charts - on a non-identified, aggregated,
anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
--
Guy Ellis
www.traverse.com.au
T: +61 3 9386 4435 M: +61 419 398 234
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
--
Guy Ellis
***@traverse.com.au
www.traverse.com.au
T: +61 3 9386 4435 M: +61 419 398 234
Beeson, Ayden
2013-09-12 22:10:03 UTC
Permalink
However, the bandwidth drop is even more drastic on vdsl. At a distance of even 500 metres your average bandwidth is drastically lower then where it started.

Loading Image...

Unfortunately, as you can see by the graph, you have to be less than 300 metres (poor copper condition not factored in) to get the promised 50mbit and at ~500 metres+ it's worse than adsl 2+ across the board.

Guy Ellis <***@traverse.com.au> wrote:



Paul,

In response to your challenge (Exercise for the reader - work out how VDSL2 would be any different)...

In contrast to the current ADSL2+ network, there are 3 big differences with the proposed VDSL2 FTTN deployment -
(i) shorter loop lengths (700-800m)
(ii) vectoring (crosstalk--)
(iii) bonding (speed++)

While such a VDSL2 network is not as good as fibre, it's no where near as bad as the current ADSL2+ network.
Right now some poor folks are on 6km loop lengths, there's plenty of crosstalk and getting bonding working is a challenge.

Regards,
- Guy.




On 12/09/2013 12:17 PM, Paul Brooks wrote:
A recent Ofcom (UK) report has a very interesting chart of ADSL2+ line speeds:
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
Figure 4 on page 11:
[cid:***@traverse.com.au]
This chart is effectively the result of the ADSL line-sync/attenuation curve combined with the increasing area of circles of increasing radius around the exchange - and demonstrates very clearly why so many people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)

Also scatter-plots of sync-speed with line-length, as per Figure 8 from another UK report:

[cid:***@traverse.com.au]
Now every DSLAM network operator can put together similar charts - but I'm not aware of any stats for Australian networks, apart from the heat maps put out by iiNet and the adsl2exchanges.com.au site, which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting together similar charts for the Australian networks, to see how our copper loop network varies from the UK network. If anyone is willing to share data or statistics, I'm very interested in pulling together similar Australian charts - on a non-identified, aggregated, anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.






_______________________________________________
AusNOG mailing list
***@lists.ausnog.net<mailto:***@lists.ausnog.net>
http://lists.ausnog.net/mailman/listinfo/ausnog




--
Guy Ellis
***@traverse.com.au<mailto:***@traverse.com.au>
www.traverse.com.au<http://www.traverse.com.au>
T: +61 3 9386 4435 M: +61 419 398 234


[cid:csu-logo3e61.bmp]<http://www.csu.edu.au/>

| ALBURY-WODONGA | BATHURST | CANBERRA | DUBBO | GOULBURN | MELBOURNE | ONTARIO | ORANGE | PORT MACQUARIE | SYDNEY | WAGGA WAGGA |

________________________________
LEGAL NOTICE
This email (and any attachment) is confidential and is intended for the use of the addressee(s) only. If you are not the intended recipient of this email, you must not copy, distribute, take any action in reliance on it or disclose it to anyone. Any confidentiality is not waived or lost by reason of mistaken delivery. Email should be checked for viruses and defects before opening. Charles Sturt University (CSU) does not accept liability for viruses or any consequence which arise as a result of this email transmission. Email communications with CSU may be subject to automated email filtering, which could result in the delay or deletion of a legitimate email before it is read at CSU. The views expressed in this email are not necessarily those of CSU.

Charles Sturt University in Australia<http://www.csu.edu.au> The Grange Chancellery, Panorama Avenue, Bathurst NSW Australia 2795 (ABN: 83 878 708 551; CRICOS Provider Numbers: 00005F (NSW), 01947G (VIC), 02960B (ACT)). TEQSA Provider Number: PV12018
Charles Sturt University in Ontario<http://www.charlessturt.ca/> 860 Harrington Court, Burlington Ontario Canada L7N 3N4 Registration: www.peqab.ca<http://www.peqab.ca>

Consider the environment before printing this email.

Disclaimer added by CodeTwo Exchange Rules 2007
www.codetwo.com<http://www.codetwo.com>
Paul Brooks
2013-09-12 23:33:10 UTC
Permalink
Post by Beeson, Ayden
Unfortunately, as you can see by the graph, you have to be less than 300 metres (poor copper condition not factored in) to get the promised 50mbit and at ~500 metres+ it's worse than adsl 2+ across the board.
VDSL2 is really just ADSL2+++++. All chipsets I know of will gracefully transition
from VDSL2 to ADSL2+ or whichever system provides the highest capacity, so I doubt
anyone will get *worse* than they would with ADSL2+. At worst it should fall back to
ADSL2+ mode.
P.
Jeremy Visser
2013-09-14 03:57:42 UTC
Permalink
Post by Paul Brooks
VDSL2 is really just ADSL2+++++. All chipsets I know of will
gracefully transition from VDSL2 to ADSL2+ or whichever system
provides the highest capacity, so I doubt anyone will get*worse*
than they would with ADSL2+. At worst it should fall back to ADSL2+
mode.
On the other hand, I've had customers who have moved from ADSL1 syncing
at a full 1536 Kbit/s downstream to ADSL2+ which managed less than 1000
Kbit/s on their long lines.

Based on that, I would guess there are instances where VDSL2 in turn
performs worse than ADSL2+.
John Edwards
2013-09-14 05:05:00 UTC
Permalink
Hi Jeremy,
On the other hand, I've had customers who have moved from ADSL1 syncing at a full 1536 Kbit/s downstream to ADSL2+ which managed less than 1000 Kbit/s on their long lines.
Based on that, I would guess there are instances where VDSL2 in turn
performs worse than ADSL2+.
This is probably more to do with DSLAM settings and profile tuning rather than a specific ADSL2 fault.

VDSL2 will bring with it a default of Packet-Transfer-Mode rather than ATM as the framing method. This alone is going to bring a 10% speed boost. A few more percentage points can be picked up by using IPoE rather than PPPoE at the higher layer.

If Vectoring is used, there's a fair chance that those with long lines will get better access to reliable low-frequency carriers while their neighbours with better propinquity utilise higher frequencies.
At worst it should fall back to ADSL2+ mode.
This assumes that the solution can get the PTM/ATM fallback working right for older modems.

Also note that vectoring is going to negate the ability to "fall back" because everyone will need vectoring-compatible CPE.

John
Paul Brooks
2013-09-15 08:26:10 UTC
Permalink
Post by John Edwards
At worst it should fall back to ADSL2+ mode.
This assumes that the solution can get the PTM/ATM fallback working right for older modems.
Also note that vectoring is going to negate the ability to "fall back" because everyone will need vectoring-compatible CPE.
Not really - lines that fall back to ADSL2+ mode will only be using frequencies up to
2.2 MHz, and probably only up to about 1.5 MHz or so at those distances. The higher
VDSL2 frequencies used by other shorter lines aren't bothered by them.

Anyway, at least 2 vendors claim they can compensate for non-vectoring-compatible
VDSL2 CPE by varying channel maps and transmission powers at the DSLAM for those
lines, and achieve almost the same results.
I'd still really like to see it to confirm, but thats what the brochureware claims at
least.

P.
Guy Ellis
2013-09-12 23:33:40 UTC
Permalink
Hi Ayden,

Do you know what VDSL2 profile was used for those tests?
It doesn't look anything like our results or AT&T's for that matter.

Normally VDSL2 and ADSL2+ intersect at around the 1.6km mark!

Regards,
- GUy.
Post by Beeson, Ayden
However, the bandwidth drop is even more drastic on vdsl. At a distance of even 500 metres your average bandwidth is drastically lower then where it started.
http://nbnmyths.files.wordpress.com/2011/06/fttn-speed-graph.gif
Unfortunately, as you can see by the graph, you have to be less than 300 metres (poor copper condition not factored in) to get the promised 50mbit and at ~500 metres+ it's worse than adsl 2+ across the board.
Paul,
In response to your challenge (Exercise for the reader - work out how VDSL2 would be any different)...
In contrast to the current ADSL2+ network, there are 3 big differences with the proposed VDSL2 FTTN deployment -
(i) shorter loop lengths (700-800m)
(ii) vectoring (crosstalk--)
(iii) bonding (speed++)
While such a VDSL2 network is not as good as fibre, it's no where near as bad as the current ADSL2+ network.
Right now some poor folks are on 6km loop lengths, there's plenty of crosstalk and getting bonding working is a challenge.
Regards,
- Guy.
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
This chart is effectively the result of the ADSL line-sync/attenuation curve combined with the increasing area of circles of increasing radius around the exchange - and demonstrates very clearly why so many people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)
Now every DSLAM network operator can put together similar charts - but I'm not aware of any stats for Australian networks, apart from the heat maps put out by iiNet and the adsl2exchanges.com.au site, which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting together similar charts for the Australian networks, to see how our copper loop network varies from the UK network. If anyone is willing to share data or statistics, I'm very interested in pulling together similar Australian charts - on a non-identified, aggregated, anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
--
Guy Ellis
www.traverse.com.au<http://www.traverse.com.au>
T: +61 3 9386 4435 M: +61 419 398 234
[cid:csu-logo3e61.bmp]<http://www.csu.edu.au/>
| ALBURY-WODONGA | BATHURST | CANBERRA | DUBBO | GOULBURN | MELBOURNE | ONTARIO | ORANGE | PORT MACQUARIE | SYDNEY | WAGGA WAGGA |
________________________________
LEGAL NOTICE
This email (and any attachment) is confidential and is intended for the use of the addressee(s) only. If you are not the intended recipient of this email, you must not copy, distribute, take any action in reliance on it or disclose it to anyone. Any confidentiality is not waived or lost by reason of mistaken delivery. Email should be checked for viruses and defects before opening. Charles Sturt University (CSU) does not accept liability for viruses or any consequence which arise as a result of this email transmission. Email communications with CSU may be subject to automated email filtering, which could result in the delay or deletion of a legitimate email before it is read at CSU. The views expressed in this email are not necessarily those of CSU.
Charles Sturt University in Australia<http://www.csu.edu.au> The Grange Chancellery, Panorama Avenue, Bathurst NSW Australia 2795 (ABN: 83 878 708 551; CRICOS Provider Numbers: 00005F (NSW), 01947G (VIC), 02960B (ACT)). TEQSA Provider Number: PV12018
Charles Sturt University in Ontario<http://www.charlessturt.ca/> 860 Harrington Court, Burlington Ontario Canada L7N 3N4 Registration: www.peqab.ca<http://www.peqab.ca>
Consider the environment before printing this email.
Disclaimer added by CodeTwo Exchange Rules 2007
www.codetwo.com<http://www.codetwo.com>
--
Guy Ellis
***@traverse.com.au
www.traverse.com.au
T: +61 3 9386 4435 M: +61 419 398 234
Beeson, Ayden
2013-09-12 23:36:09 UTC
Permalink
Are the dslam's capable of that without special configuration too? That's pretty cool, though if they build it the way they are "guaranteeing" it shouldn't ever have to...
Post by Beeson, Ayden
Unfortunately, as you can see by the graph, you have to be less than 300 metres (poor copper condition not factored in) to get the promised 50mbit and at ~500 metres+ it's worse than adsl 2+ across the board.
VDSL2 is really just ADSL2+++++. All chipsets I know of will gracefully transition
from VDSL2 to ADSL2+ or whichever system provides the highest capacity, so I doubt
anyone will get *worse* than they would with ADSL2+. At worst it should fall back to
ADSL2+ mode.
P.

_______________________________________________
AusNOG mailing list
***@lists.ausnog.net
http://lists.ausnog.net/mailman/listinfo/ausnog
Charles Sturt University

| ALBURY-WODONGA | BATHURST | CANBERRA | DUBBO | GOULBURN | MELBOURNE | ONTARIO | ORANGE | PORT MACQUARIE | SYDNEY | WAGGA WAGGA |

LEGAL NOTICE
This email (and any attachment) is confidential and is intended for the use of the addressee(s) only. If you are not the intended recipient of this email, you must not copy, distribute, take any action in reliance on it or disclose it to anyone. Any confidentiality is not waived or lost by reason of mistaken delivery. Email should be checked for viruses and defects before opening. Charles Sturt University (CSU) does not accept liability for viruses or any consequence which arise as a result of this email transmission. Email communications with CSU may be subject to automated email filtering, which could result in the delay or deletion of a legitimate email before it is read at CSU. The views expressed in this email are not necessarily those of CSU.

Charles Sturt University in Australia http://www.csu.edu.au The Grange Chancellery, Panorama Avenue, Bathurst NSW Australia 2795 (ABN: 83 878 708 551; CRICOS Provider Numbers: 00005F (NSW), 01947G (VIC), 02960B (ACT)). TEQSA Provider Number: PV12018

Charles Sturt University in Ontario http://www.charlessturt.ca 860 Harrington Court, Burlington Ontario Canada L7N 3N4 Registration: www.peqab.ca

Consider the environment before printing this email.

Disclaimer added by CodeTwo Exchange Rules 2007
http://www.codetwo.com
Beeson, Ayden
2013-09-12 23:39:49 UTC
Permalink
Absolutely no idea tbh, it's an ofcom graph, that's all I know.

Seemed optimistic for adsl2+ as well tbh... I would have assumed closer to 1-1.5km for the crossover as well...

Guy Ellis <***@traverse.com.au> wrote:


Hi Ayden,

Do you know what VDSL2 profile was used for those tests?
It doesn't look anything like our results or AT&T's for that matter.

Normally VDSL2 and ADSL2+ intersect at around the 1.6km mark!

Regards,
- GUy.
Post by Beeson, Ayden
However, the bandwidth drop is even more drastic on vdsl. At a distance of even 500 metres your average bandwidth is drastically lower then where it started.
http://nbnmyths.files.wordpress.com/2011/06/fttn-speed-graph.gif
Unfortunately, as you can see by the graph, you have to be less than 300 metres (poor copper condition not factored in) to get the promised 50mbit and at ~500 metres+ it's worse than adsl 2+ across the board.
Paul,
In response to your challenge (Exercise for the reader - work out how VDSL2 would be any different)...
In contrast to the current ADSL2+ network, there are 3 big differences with the proposed VDSL2 FTTN deployment -
(i) shorter loop lengths (700-800m)
(ii) vectoring (crosstalk--)
(iii) bonding (speed++)
While such a VDSL2 network is not as good as fibre, it's no where near as bad as the current ADSL2+ network.
Right now some poor folks are on 6km loop lengths, there's plenty of crosstalk and getting bonding working is a challenge.
Regards,
- Guy.
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
This chart is effectively the result of the ADSL line-sync/attenuation curve combined with the increasing area of circles of increasing radius around the exchange - and demonstrates very clearly why so many people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)
Now every DSLAM network operator can put together similar charts - but I'm not aware of any stats for Australian networks, apart from the heat maps put out by iiNet and the adsl2exchanges.com.au site, which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting together similar charts for the Australian networks, to see how our copper loop network varies from the UK network. If anyone is willing to share data or statistics, I'm very interested in pulling together similar Australian charts - on a non-identified, aggregated, anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
--
Guy Ellis
www.traverse.com.au<http://www.traverse.com.au>
T: +61 3 9386 4435 M: +61 419 398 234
[cid:csu-logo3e61.bmp]<http://www.csu.edu.au/>
| ALBURY-WODONGA | BATHURST | CANBERRA | DUBBO | GOULBURN | MELBOURNE | ONTARIO | ORANGE | PORT MACQUARIE | SYDNEY | WAGGA WAGGA |
________________________________
LEGAL NOTICE
This email (and any attachment) is confidential and is intended for the use of the addressee(s) only. If you are not the intended recipient of this email, you must not copy, distribute, take any action in reliance on it or disclose it to anyone. Any confidentiality is not waived or lost by reason of mistaken delivery. Email should be checked for viruses and defects before opening. Charles Sturt University (CSU) does not accept liability for viruses or any consequence which arise as a result of this email transmission. Email communications with CSU may be subject to automated email filtering, which could result in the delay or deletion of a legitimate email before it is read at CSU. The views expressed in this email are not necessarily those of CSU.
Charles Sturt University in Australia<http://www.csu.edu.au> The Grange Chancellery, Panorama Avenue, Bathurst NSW Australia 2795 (ABN: 83 878 708 551; CRICOS Provider Numbers: 00005F (NSW), 01947G (VIC), 02960B (ACT)). TEQSA Provider Number: PV12018
Charles Sturt University in Ontario<http://www.charlessturt.ca/> 860 Harrington Court, Burlington Ontario Canada L7N 3N4 Registration: www.peqab.ca<http://www.peqab.ca>
Consider the environment before printing this email.
Disclaimer added by CodeTwo Exchange Rules 2007
www.codetwo.com<http://www.codetwo.com>
--
Guy Ellis
***@traverse.com.au
www.traverse.com.au
T: +61 3 9386 4435 M: +61 419 398 234

Charles Sturt University

| ALBURY-WODONGA | BATHURST | CANBERRA | DUBBO | GOULBURN | MELBOURNE | ONTARIO | ORANGE | PORT MACQUARIE | SYDNEY | WAGGA WAGGA |

LEGAL NOTICE
This email (and any attachment) is confidential and is intended for the use of the addressee(s) only. If you are not the intended recipient of this email, you must not copy, distribute, take any action in reliance on it or disclose it to anyone. Any confidentiality is not waived or lost by reason of mistaken delivery. Email should be checked for viruses and defects before opening. Charles Sturt University (CSU) does not accept liability for viruses or any consequence which arise as a result of this email transmission. Email communications with CSU may be subject to automated email filtering, which could result in the delay or deletion of a legitimate email before it is read at CSU. The views expressed in this email are not necessarily those of CSU.

Charles Sturt University in Australia http://www.csu.edu.au The Grange Chancellery, Panorama Avenue, Bathurst NSW Australia 2795 (ABN: 83 878 708 551; CRICOS Provider Numbers: 00005F (NSW), 01947G (VIC), 02960B (ACT)). TEQSA Provider Number: PV12018

Charles Sturt University in Ontario http://www.charlessturt.ca 860 Harrington Court, Burlington Ontario Canada L7N 3N4 Registration: www.peqab.ca

Consider the environment before printing this email.

Disclaimer added by CodeTwo Exchange Rules 2007
http://www.codetwo.com
Tony
2013-09-12 23:55:26 UTC
Permalink
You mean like this (hosted on the same website):
Loading Image...

or this:
Loading Image...


The other thing I'm curious about on that original graph is whether it is downstream, upstream or both combined ?

If at 2000m you get either 18/1Mbps ADSL2+ or 11/11Mbps VDSL2 then thats
fairly close to the same amount of total bandwidth (19 v's 22). This probably doesn't please the user though as they
want faster downloads (to "obtain" the latest TV eps) and hence an
option to choose which "DSL" you want might be of benefit depending on your intended usage of the link.


I'm not arguing for a FTTN network (I want my FTTH) but discussion needs to be responsible.


regards,
Tony.
Post by Paul Brooks
________________________________
Sent: Friday, 13 September 2013 9:39 AM
Subject: Re: [AusNOG] ADSL2+ line sync data
Absolutely no idea tbh, it's an ofcom graph, that's all I know.
Seemed optimistic for adsl2+ as well tbh... I would have assumed closer to 1-1.5km for the crossover as well...
Hi Ayden,
Do you know what VDSL2 profile was used for those tests?
It doesn't look anything like our results or AT&T's for that matter.
Normally VDSL2 and ADSL2+ intersect at around the 1.6km mark!
Regards,
  - GUy.
Post by Beeson, Ayden
However, the bandwidth drop is even more drastic on vdsl. At a distance of even 500 metres your average bandwidth is drastically lower then where it started.
http://nbnmyths.files.wordpress.com/2011/06/fttn-speed-graph.gif
Unfortunately, as you can see by the graph, you have to be less than 300 metres (poor copper condition not factored in) to get the promised 50mbit and at ~500 metres+ it's worse than adsl 2+ across the board.
Paul,
In response to your challenge (Exercise for the reader - work out how VDSL2 would be any different)...
In contrast to the current ADSL2+ network, there are 3 big differences with the proposed VDSL2 FTTN deployment -
(i) shorter loop lengths (700-800m)
(ii) vectoring (crosstalk--)
(iii) bonding (speed++)
While such a VDSL2 network is not as good as fibre, it's no where near as bad as the current ADSL2+ network.
Right now some poor folks are on 6km loop lengths, there's plenty of crosstalk and getting bonding working is a challenge.
Regards,
  - Guy.
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
This chart is effectively the result of the ADSL line-sync/attenuation curve combined with the increasing area of circles of increasing radius around the exchange - and demonstrates very clearly why so many people get low ADSL2+ line speeds.
(Exercise for the reader - work out how VDSL2 would be any different)
Now every DSLAM network operator can put together similar charts - but I'm not aware of any stats for Australian networks, apart from the heat maps put out by iiNet and the adsl2exchanges.com.au site, which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting together similar charts for the Australian networks, to see how our copper loop network varies from the UK network. If anyone is willing to share data or statistics, I'm very interested in pulling together similar Australian charts - on a non-identified, aggregated, anonymised basis if you wish.
Please contact me off-list - thanks.
Paul.
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
--
Guy Ellis
www.traverse.com.au<http://www.traverse.com.au>
T: +61 3 9386 4435 M: +61 419 398 234
[cid:csu-logo3e61.bmp]<http://www.csu.edu.au/>
|  ALBURY-WODONGA  |  BATHURST  |  CANBERRA  |  DUBBO  |  GOULBURN  |  MELBOURNE  |  ONTARIO  |  ORANGE  |  PORT MACQUARIE  |  SYDNEY  |  WAGGA WAGGA  |
________________________________
LEGAL NOTICE
This email (and any attachment) is confidential and is intended for the use of the addressee(s) only. If you are not the intended recipient of this email, you must not copy, distribute, take any action in reliance on it or disclose it to anyone. Any confidentiality is not waived or lost by reason of mistaken delivery. Email should be checked for viruses and defects before opening. Charles Sturt University (CSU) does not accept liability for viruses or any consequence which arise as a result of this email transmission. Email communications with CSU may be subject to automated email filtering, which could result in the delay or deletion of a legitimate email before it is read at CSU. The views expressed in this email are not necessarily those of CSU.
Charles Sturt University in Australia<http://www.csu.edu.au> The Grange Chancellery, Panorama Avenue, Bathurst NSW Australia 2795 (ABN: 83 878 708 551; CRICOS Provider Numbers: 00005F (NSW), 01947G (VIC), 02960B (ACT)). TEQSA Provider Number: PV12018
Charles Sturt University in Ontario<http://www.charlessturt.ca/> 860 Harrington Court, Burlington Ontario Canada L7N 3N4 Registration: www.peqab.ca<http://www.peqab.ca>
Consider the environment before printing this email.
Disclaimer added by CodeTwo Exchange Rules 2007
www.codetwo.com<http://www.codetwo.com>
--
Guy Ellis
www.traverse.com.au
T: +61 3 9386 4435 M: +61 419 398 234
Charles Sturt University
| ALBURY-WODONGA | BATHURST | CANBERRA | DUBBO | GOULBURN | MELBOURNE | ONTARIO | ORANGE | PORT MACQUARIE | SYDNEY | WAGGA WAGGA |
LEGAL NOTICE
This email (and any attachment) is confidential and is intended for the use of the addressee(s) only. If you are not the intended recipient of this email, you must not copy, distribute, take any action in reliance on it or disclose it to anyone. Any confidentiality is not waived or lost by reason of mistaken delivery. Email should be checked for viruses and defects before opening. Charles Sturt University (CSU) does not accept liability for viruses or any consequence which arise as a result of this email transmission. Email communications with CSU may be subject to automated email filtering, which could result in the delay or deletion of a legitimate email before it is read at CSU. The views expressed in this email are not necessarily those of CSU.
Charles Sturt University in Australia  http://www.csu.edu.au The Grange Chancellery, Panorama Avenue, Bathurst NSW Australia 2795  (ABN: 83 878 708 551; CRICOS Provider Numbers: 00005F (NSW), 01947G (VIC), 02960B (ACT)). TEQSA Provider Number: PV12018
Charles Sturt University in Ontario  http://www.charlessturt.ca 860 Harrington Court, Burlington Ontario Canada L7N 3N4  Registration: www.peqab.ca
Consider the environment before printing this email.
Disclaimer added by CodeTwo Exchange Rules 2007
http://www.codetwo.com
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Mark ZZZ Smith
2013-09-13 11:41:35 UTC
Permalink
Post by Paul Brooks
________________________________
Sent: Friday, 13 September 2013 9:55 AM
Subject: Re: [AusNOG] ADSL2+ line sync data
http://nbnmyths.files.wordpress.com/2010/09/dsl-speed-comparisons.jpg
http://www.asotel.net/files/pics/news/grafikVDSL2%29eng.png
The other thing I'm curious about on that original graph is whether it is downstream, upstream or both combined ?
If at 2000m you get either 18/1Mbps ADSL2+ or 11/11Mbps VDSL2 then thats
fairly close to the same amount of total bandwidth (19 v's 22). This probably doesn't please the user though as they
want faster downloads (to "obtain" the latest TV eps) and hence an
option to choose which "DSL" you want might be of benefit depending on your intended usage of the link.
Sad thing about this is that doing that appears to be a good idea, but that actual best thing for performance is symmetry, due to how TCP uses ACK feedback to determine it's send rate. Congestion on the uplink, which is more likely to occur when the uplink bandwidth is lower than downlink bandwidth, will slow down downloads:

BCP69/RFC3449, "TCP Performance Implications of Network Path Asymmetry"
http://tools.ietf.org/html/rfc3449
Post by Paul Brooks
I'm not arguing for a FTTN network (I want my FTTH) but discussion needs to be responsible.
regards,
Tony.
Post by Paul Brooks
________________________________
Sent: Friday, 13 September 2013 9:39 AM
Subject: Re: [AusNOG] ADSL2+ line sync data
Absolutely no idea tbh, it's an ofcom graph, that's all I know.
Seemed optimistic for adsl2+ as well tbh... I would have assumed closer to 1-1.5km for the crossover as well...
Guy Ellis
Hi Ayden,
Do you know what VDSL2 profile was used for those tests?
It doesn't look anything like our results or AT&T's for that matter.
Normally VDSL2 and ADSL2+ intersect at around the 1.6km mark!
Regards,
  - GUy.
Post by Beeson, Ayden
However, the bandwidth drop is even more drastic on vdsl. At a distance of even 500 metres your average bandwidth is drastically lower then where it started.
http://nbnmyths.files.wordpress.com/2011/06/fttn-speed-graph.gif
Unfortunately, as you can see by the graph, you have to be less than 300 metres (poor copper condition not factored in) to get the promised 50mbit and at ~500 metres+ it's
worse than adsl 2+ across the board.
Post by Paul Brooks
Post by Paul Brooks
Post by Beeson, Ayden
Paul,
In response to your challenge (Exercise for the reader - work out how VDSL2 would be any different)...
In contrast to the current ADSL2+ network, there are 3 big differences with the proposed VDSL2 FTTN deployment -
(i) shorter loop lengths (700-800m)
(ii) vectoring (crosstalk--)
(iii) bonding (speed++)
While such a VDSL2 network is not as good as fibre, it's no where near as bad as the current ADSL2+ network.
Right now some poor folks are on 6km loop lengths, there's plenty of crosstalk and getting bonding working is a challenge.
Regards,
   - Guy.
On 12/09/2013 12:17 PM, Paul Brooks
Ofcom Infrastructure Report 2012 Update
http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/broadband-speeds/infrastructure-report-2012/
http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/infrastructure-report/Infrastructure-report2012.pdf
This chart is effectively the result of the ADSL
line-sync/attenuation curve combined with the increasing area of circles of increasing radius around the exchange - and demonstrates very clearly why so many people get low ADSL2+ line speeds.
Post by Paul Brooks
Post by Paul Brooks
Post by Beeson, Ayden
(Exercise for the reader - work out how VDSL2 would be any different)
Now every DSLAM network operator can put together similar charts - but I'm not aware of any stats for Australian networks, apart from the heat maps put out by iiNet and the adsl2exchanges.com.au site, which aren't quite what I'm looking for.
For all you DSLAM operators - I would be very interested in putting together similar charts for the Australian networks, to see how our copper loop
network varies from the UK network. If anyone is willing to share data or statistics, I'm very interested in pulling together similar Australian charts - on a non-identified, aggregated, anonymised basis if you wish.
Post by Paul Brooks
Post by Paul Brooks
Post by Beeson, Ayden
Please contact me off-list - thanks.
Paul.
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
--
Guy Ellis
www.traverse.com.au<http://www.traverse.com.au>
T: +61 3 9386 4435 M: +61 419 398 234
[cid:csu-logo3e61.bmp]<http://www.csu.edu.au/>
|   ALBURY-WODONGA   |   BATHURST   |   CANBERRA   |   DUBBO   |   GOULBURN   |   MELBOURNE   |   ONTARIO   |   ORANGE   |   PORT MACQUARIE   |   SYDNEY   |   WAGGA WAGGA   |
________________________________
LEGAL NOTICE
This email (and any attachment) is confidential and is intended for the use of the addressee(s) only. If you are not the intended recipient of this email, you must not copy, distribute, take
any action in reliance on it or disclose it to anyone. Any confidentiality is not waived or lost by reason of mistaken delivery. Email should be checked for viruses and defects before opening. Charles Sturt University (CSU) does not accept liability for viruses or any consequence which arise as a result of this email transmission. Email communications with CSU may be subject to automated email filtering, which could result in the delay or deletion of a legitimate email before it is read at CSU. The views expressed in this email are not necessarily those of CSU.
Post by Paul Brooks
Post by Paul Brooks
Post by Beeson, Ayden
Charles Sturt University in Australia<http://www.csu.edu.au> The Grange Chancellery, Panorama Avenue, Bathurst NSW Australia 2795 (ABN: 83 878 708 551; CRICOS Provider Numbers: 00005F (NSW), 01947G (VIC), 02960B (ACT)). TEQSA Provider Number: PV12018
Charles Sturt University in Ontario<http://www.charlessturt.ca/> 860 Harrington Court, Burlington Ontario Canada L7N 3N4 Registration: www.peqab.ca<http://www.peqab.ca>
Consider the environment before printing this email.
Disclaimer added by CodeTwo Exchange Rules 2007
www.codetwo.com<http://www.codetwo.com>
--
Guy Ellis
www.traverse.com.au
T: +61 3 9386 4435 M: +61 419 398 234
Charles Sturt University
| ALBURY-WODONGA | BATHURST | CANBERRA | DUBBO | GOULBURN | MELBOURNE | ONTARIO | ORANGE | PORT MACQUARIE | SYDNEY | WAGGA WAGGA |
LEGAL NOTICE
This email (and any attachment) is confidential and is intended for the use
of the addressee(s) only. If you are not the intended recipient of this email, you must not copy, distribute, take any action in reliance on it or disclose it to anyone. Any confidentiality is not waived or lost by reason of mistaken delivery. Email should be checked for viruses and defects before opening. Charles Sturt University (CSU) does not accept liability for viruses or any consequence which arise as a result of this email transmission. Email communications with CSU may be subject to automated email filtering, which could result in the delay or deletion of a legitimate email before it is read at CSU. The views expressed in this email are not necessarily those of CSU.
Post by Paul Brooks
Post by Paul Brooks
Charles Sturt University in Australia  http://www.csu.edu.au  The Grange Chancellery, Panorama Avenue, Bathurst NSW Australia 2795  (ABN: 83 878 708 551; CRICOS Provider Numbers: 00005F (NSW), 01947G (VIC), 02960B (ACT)). TEQSA Provider Number: PV12018
Charles Sturt University in Ontario  http://www.charlessturt.ca 860 Harrington Court, Burlington Ontario Canada L7N 3N4  Registration: www.peqab.ca
Consider the environment before printing this email.
Disclaimer added by CodeTwo Exchange Rules 2007
http://www.codetwo.com
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Mark ZZZ Smith
2013-09-13 11:47:32 UTC
Permalink
----- Original Message -----
Sent: Friday, 13 September 2013 9:41 PM
Subject: Re: [AusNOG] ADSL2+ line sync data
Post by Paul Brooks
Sent: Friday, 13 September 2013 9:55 AM
<snip>
Post by Paul Brooks
If at 2000m you get either 18/1Mbps ADSL2+ or 11/11Mbps VDSL2 then thats
fairly close to the same amount of total bandwidth (19 v's 22). This
probably doesn't please the user though as they
want faster downloads (to "obtain" the latest TV eps) and hence an
option to choose which "DSL" you want might be of benefit depending on
your intended usage of the link.
Sad thing about this is that doing that appears to be a good idea, but that
actual best thing for performance is symmetry, due to how TCP uses ACK feedback
to determine it's send rate. Congestion on the uplink, which is more likely
to occur when the uplink bandwidth is lower than downlink bandwidth, will slow
BCP69/RFC3449, "TCP Performance Implications of Network Path
Asymmetry"
http://tools.ietf.org/html/rfc3449
I should add that Google gets it, which is why their FTTP is symmetric.

http://googlefiberblog.blogspot.com.au/2012/04/construction-update.html


Q: I see a label on your diagram, “Gigabit Symmetric Fiber Connectivity.” What does that mean?
Joshua D'Alton
2013-09-13 12:25:59 UTC
Permalink
That RFC isn't very accurate/applicable any more, the bandwidths we're
talking about now throws most of those calculations and factors etc out of
the window. The maths is still solid, but the results are ... not. Stream
delay (buffer size, mean RTT for a single TCP packet etc) now being 100th
of what it was with dialup, and so on.

Google might have that Q and A, and they are correct in theory that it will
let you upload alot as well (whatever that means *caugh P2P caugh*), but
they needn't have gone anywhere near symmetrical to rule out 99.999% risk
of downloads being slowed due to upload congestion. Put another way,
looking at that 18/1 vs 11/11 the majority of users will be experiencing
downstream congestion before they see upstream congestion.

I don't know exactly what ratio is 'ideal', especially given a swath of
difference user types, but if we were to maximise bandwidth usage and
'altruistically' determine the rates, it definitely wouldn't be 1:1
(symmetrical), perhaps not 18:1, but definitely somewhere inbetween given
total bandwidth > 10Mbps or so (haven't sat down and figured it out
exactly, but that is probably ballpark given tcp overhead figures, ie for
the download stream, and then whatever extra upload data is being sent).

And then that's not to mention fiber typically 'has to be' symmetrical so
theres really no administrative difference, or in fact its less burden
leaving as is vs upload throttling (for whatever reason one might want to
do that, and if that was even an issue).




On Fri, Sep 13, 2013 at 9:47 PM, Mark ZZZ Smith
Post by Mark ZZZ Smith
----- Original Message -----
Sent: Friday, 13 September 2013 9:41 PM
Subject: Re: [AusNOG] ADSL2+ line sync data
Post by Paul Brooks
Sent: Friday, 13 September 2013 9:55 AM
<snip>
Post by Paul Brooks
If at 2000m you get either 18/1Mbps ADSL2+ or 11/11Mbps VDSL2 then thats
fairly close to the same amount of total bandwidth (19 v's 22). This
probably doesn't please the user though as they
want faster downloads (to "obtain" the latest TV eps) and hence an
option to choose which "DSL" you want might be of benefit depending on
your intended usage of the link.
Sad thing about this is that doing that appears to be a good idea, but
that
actual best thing for performance is symmetry, due to how TCP uses ACK
feedback
to determine it's send rate. Congestion on the uplink, which is more
likely
to occur when the uplink bandwidth is lower than downlink bandwidth,
will slow
BCP69/RFC3449, "TCP Performance Implications of Network Path
Asymmetry"
http://tools.ietf.org/html/rfc3449
I should add that Google gets it, which is why their FTTP is symmetric.
http://googlefiberblog.blogspot.com.au/2012/04/construction-update.html
Q: I see a label on your diagram, “Gigabit Symmetric Fiber Connectivity.”
What does that mean?
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Mark ZZZ Smith
2013-09-14 00:15:24 UTC
Permalink
Post by Paul Brooks
________________________________
Sent: Friday, 13 September 2013 10:25 PM
Subject: Re: [AusNOG] ADSL2+ line sync data
That RFC isn't very accurate/applicable any more, the bandwidths we're talking about now throws most of those calculations and factors etc out of the window. The maths is still solid, but the results are ... not. Stream delay (buffer size, mean RTT for a single TCP packet etc) now being 100th of what it was with dialup, and so on.
It's still an IETF Best Current Practice. The IETF will deprecate BCPs if they don't apply any more. Given that that are over 6000 RFCs, and only 183 BCPs, RFCs are not given BCP status without significant consideration.


If you think it only applies to dialup, then I don't think you've read it.

For example, from the appendix,

     "Most data over cable TV networks (e.g., DOCSIS [ITU01, DS00]),
      where the analogue channels assigned for upstream communication
      (i.e., in the reverse direction) are narrower and may be more
      noisy than those assigned for the downstream link.  As a
      consequence, the upstream and downstream links differ in their
      transmission rate. For example, in DOCSIS 1.0 [DS00], the
      downstream transmission rate is either 27 or 52 Mbps.  Upstream
      transmission rates may be dynamically selected to be one of a
      series of rates which range between 166 kbps to 9 Mbps."

A broadband speed of DOCSIS best case 52Mbps/9Mbps is certainly not dialup.
Post by Paul Brooks
Google might have that Q and A, and they are correct in theory that it will let you upload alot as well (whatever that means *caugh P2P caugh*), but they needn't have gone anywhere near symmetrical to rule out 99.999% risk of downloads being slowed due to upload congestion. Put another way, looking at that 18/1 vs 11/11 the majority of users will be experiencing downstream congestion before they see upstream congestion.
It's the ratio of downstream to update bandwidth that matters, and the likelihood of congestion in the upstream direction, not so much the bandwidth involved. The greater the ratio of downstream to upstream bandwidth, the more likely the problem is going to occur.


A single TCP stream is less likely to see this issue, as there is very little competition for uplink bandwidth that can cause ACKs to be queued and therefore delayed. However, have a large upload going (e.g. video upload to youtube) while also having a large download going (e.g., watching a youtube video), and you're more likely to see the download effected because of ACK queuing and therefore delaying occurring on the uplink.

Symmetry doesn't eliminate congestion (nothing does), but it can provide equal performance and quality in both directions for TCP and other protocols and applications. Video can be the same quality in both directions, not better in one direction than the other (asymmetry of bandwidth is going to put a real cramp on telemedicine, as the patient will see the doctor really well, but the doctor won't see the patient really well - and doctors won't accept that because their malpractice premiums would then go up). Uploads will have as much congestion effect as downloads. 
Post by Paul Brooks
I don't know exactly what ratio is 'ideal',
True, but the IETF do, and they're the experts on Internet protocols. They designed TCP, and they've published an RFC that has become a Best Common Practice on how asymmetry impacts its performance.

If you read nothing else, read the following from the summary,

"Asymmetry, and particular high asymmetry, raises a set of TCP performance issues."
Post by Paul Brooks
especially given a swath of difference user types, but if we were to maximise bandwidth usage and 'altruistically' determine the rates, it definitely wouldn't be 1:1 >(symmetrical), perhaps not 18:1, but definitely somewhere inbetween given total bandwidth > 10Mbps or so (haven't sat down and figured it out exactly, but that is >probably ballpark given tcp overhead figures, ie for the download stream, and then whatever extra upload data is being sent).
And then that's not to mention fiber typically 'has to be' symmetrical so theres really no administrative difference, or in fact its less burden leaving as is vs upload throttling (for whatever reason one might want to do that, and if that was even an issue).
Post by Mark ZZZ Smith
----- Original Message -----
Sent: Friday, 13 September 2013 9:41 PM
Subject: Re: [AusNOG] ADSL2+ line sync data
Post by Paul Brooks
Sent: Friday, 13 September 2013 9:55 AM
<snip>
Post by Paul Brooks
If at 2000m you get either 18/1Mbps ADSL2+ or 11/11Mbps VDSL2 then thats
fairly close to the same amount of total bandwidth (19 v's 22). This
probably doesn't please the user though as they
want faster downloads (to "obtain" the latest TV eps) and hence an
option to choose which "DSL" you want might be of benefit depending on
your intended usage of the link.
Sad thing about this is that doing that appears to be a good idea, but that
actual best thing for performance is symmetry, due to how TCP uses ACK feedback
to determine it's send rate. Congestion on the uplink, which is more likely
to occur when the uplink bandwidth is lower than downlink bandwidth, will slow
BCP69/RFC3449, "TCP Performance Implications of Network Path Asymmetry"
http://tools.ietf.org/html/rfc3449
I should add that Google gets it, which is why their FTTP is symmetric.
http://googlefiberblog.blogspot.com.au/2012/04/construction-update.html
Q: I see a label on your diagram, “Gigabit Symmetric Fiber Connectivity.” What does that mean?
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
grenville armitage
2013-09-14 01:22:39 UTC
Permalink
On 09/14/2013 10:15, Mark ZZZ Smith wrote:
[..]
Post by Mark ZZZ Smith
It's the ratio of downstream to update bandwidth that matters, and
the likelihood of congestion in the upstream direction, not so much
the bandwidth involved. The greater the ratio of downstream to
upstream bandwidth, the more likely the problem is going to occur.
+1

(Seriously, Mark is right about http://tools.ietf.org/html/rfc3449. Still relevant in our brave new asymmetric high speed world.)

cheers,
gja
Mark ZZZ Smith
2013-09-15 03:56:59 UTC
Permalink
----- Original Message -----
Sent: Saturday, 14 September 2013 11:22 AM
Subject: Re: [AusNOG] ADSL2+ line sync data
    [..]
Post by Mark ZZZ Smith
It's the ratio of downstream to update bandwidth that matters, and
the likelihood of congestion in the upstream direction, not so much
the bandwidth involved. The greater the ratio of downstream to
upstream bandwidth, the more likely the problem is going to occur.
+1
(Seriously, Mark is right about http://tools.ietf.org/html/rfc3449. Still
relevant in our brave new asymmetric high speed world.)
 
Where I most think it matters is that the service speeds people buy sets, quite reasonably, their expectations of what they'll be able to get out of their service.

So if a customer buys a 25/5 service for example, they'd expect that they'd be able to upload at 5Mbps while also downloading at 25Mbps. They probably won't be able to due to bandwidth asymmetry, which means that the ISP/RSP may be violating trade practices laws for false advertising. Or in the least, the ISP/RSP have to deal with customer complaints.

OTOH, if the ISP/RSP provided 25/25, then it is possible to both upload and download using TCP at full rates in both directions.

In the brave new NBN world, where people have data centers in their basement, are uploading X-rays from their iPhone 9, and streaming 4K security camera video of their cat to the Internet, I think the consequences of this asymmetry are going to become are more visible.


Regards,

Mark.
cheers,
gja
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Joshua D'Alton
2013-09-15 04:34:39 UTC
Permalink
Might be lazy sunday afternoon messing with my comprehension, but in the
former case they'll have ~4.6Mbps spare upload capacity, in the later
24.6Mbps. Neither are full rate, if I read you correctly, though in the
later case it is 98.5% line rate vs 92% which might be a noticable
difference, probably not in terms of time for something to upload (less
than 10% difference), but maybe for capability itself to upload (ie
maintaining full stream video conference or doctors etc).

But really I doubt anyone would notice, especially given speedtest doesn't
use both at the same time. In fact I think more people would complain about
actual performance (ie why don't i get 25Mbps from some remote server, or
hell even youtube/akamai/CDN cache locally), than the few that would fire
up uTorrent, download a linux iso and think "oh hey, 25mbps down, but only
4.6mbps up in the status bar, why isn't that 5mbps!?!". Not to mention
everyone already just says "up to", both for line rate reasons, and for
aforementioned remote site/connectivity bottleneck being causes for not
seeing maximum advertised speeds.

I could be wrong but I think you've shifted a bit away from ietf best
practice reasons to have symmetry, towards the fact people actually need
more upload full stop, symmetry being a happy coincidence. Which I'd agree
with, though not that symmetry is required/very.important for customer
expectation/relations reasons.

Anyway I originally said "not very applicable", and it seems we're all
talking about the fringe cases where it IS infact much more applicable, so
if we stepped back into the realm of averages and normal use cases.. Given
the choice between higher download and symmetry, the very fact that most
connections are used to consume data and not create data should and does
lead towards asymmetry and will remain that way even with datacenters in
basements and 4K cat streaming.

</arguing>


On Sun, Sep 15, 2013 at 1:56 PM, Mark ZZZ Smith
Post by Mark ZZZ Smith
So if a customer buys a 25/5 service for example, they'd expect that
they'd be able to upload at 5Mbps while also downloading at 25Mbps. They
probably won't be able to due to bandwidth asymmetry, which means that the
ISP/RSP may be violating trade practices laws for false advertising. Or in
the least, the ISP/RSP have to deal with customer complaints.
OTOH, if the ISP/RSP provided 25/25, then it is possible to both upload
and download using TCP at full rates in both directions.
Mark ZZZ Smith
2013-09-15 05:09:07 UTC
Permalink
Post by Paul Brooks
________________________________
Sent: Sunday, 15 September 2013 2:34 PM
Subject: Re: [AusNOG] ADSL2+ line sync data
Might be lazy sunday afternoon messing with my comprehension, but in the former case they'll have ~4.6Mbps spare upload capacity, in the later 24.6Mbps. Neither are full rate, if I read you correctly, though in the later case it is 98.5% line rate vs 92% which might be a noticable difference, probably not in terms of time for something to upload (less than 10% difference), but maybe for capability itself to upload (ie maintaining full stream video conference or doctors etc). 
But really I doubt anyone would notice, especially given speedtest doesn't use both at the same time. In fact I think more people would complain about actual performance (ie why don't i get 25Mbps from some remote server, or hell even youtube/akamai/CDN cache locally), than the few that would fire up uTorrent, download a linux iso and think "oh hey, 25mbps down, but only 4.6mbps up in the status bar, why isn't that 5mbps!?!". Not to mention everyone already just says "up to", both for line rate reasons, and for aforementioned remote site/connectivity bottleneck being causes for not seeing maximum advertised speeds.
Why is it acceptable to not be able provide what is being sold? Just because some people don't realise they're not getting the service they're paying for, I don't think that means it is acceptable to continue to sell a service that cannot actually be delivered.
Post by Paul Brooks
I could be wrong but I think you've shifted a bit away from ietf best practice reasons to have symmetry, towards the fact people actually need more upload full stop, symmetry being a happy coincidence.
No, I have not shifted away from the IETF BCP. People need so much "upload" that it equals their "download".

The thing you seem to be missing is it isn't the type of service or the type of user that dictates whether asymmetry will cause problems, it's whether TCP is being used or not. Since TCP is used by everybody, whether it's a residential or business service doesn't matter. *All* user or customer types will benefit from symmetry, unless they don't use TCP.

Assuming residential users will only download is ignoring the reality that people are doing things like uploading large photos and videos to facebook, flickr, youtube etc., and video conferencing with their relatives or their doctor or other professionals. If the "NBNs" of either party don't recognise this, then they're not as revolutionary or evolutionary as they and the general population think. These NBNs are going to constrain end-users to primarily being consumers of content rather than equally being possibly both consumers and producers, or being producers more than consumers.
Post by Paul Brooks
Which I'd agree with, though not that symmetry is required/very.important for customer expectation/relations reasons. 
Anyway I originally said "not very applicable", and it seems we're all talking about the fringe cases where it IS infact much more applicable, so if we stepped back into the realm of averages and normal use cases.. Given the choice between higher download and symmetry, the very fact that most connections are used to consume data and not create data should and does lead towards asymmetry and will remain that way even with datacenters in basements and 4K cat streaming.
</arguing>
Post by Mark ZZZ Smith
So if a customer buys a 25/5 service for example, they'd expect that they'd be able to upload at 5Mbps while also downloading at 25Mbps. They probably won't be able to due to bandwidth asymmetry, which means that the ISP/RSP may be violating trade practices laws for false advertising. Or in the least, the ISP/RSP have to deal with customer complaints.
OTOH, if the ISP/RSP provided 25/25, then it is possible to both upload and download using TCP at full rates in both directions.
Tony
2013-09-16 06:01:14 UTC
Permalink
________________________________
Post by Mark ZZZ Smith
The thing you seem to be missing is it isn't the type of service or the type of user that dictates whether asymmetry will cause problems, it's whether TCP is being used or not. Since TCP is used by everybody, whether it's a residential or business service doesn't matter. *All* user or customer types will benefit from symmetry, unless they don't use TCP.
Assuming residential users will only download is ignoring the reality that people are doing things like uploading large photos and videos to facebook, flickr, youtube etc., and video conferencing with their relatives or their doctor or other professionals. If the "NBNs" of either party don't recognise this, then they're not as revolutionary or evolutionary as they and the general population think. These NBNs are going to constrain end-users to primarily being consumers of content rather than equally being possibly both consumers and producers, or being producers more than consumers.
All video conf & VoIP I've seen uses UDP which means the video conferencing session to the medical professional is not going to care about ACK's at all. Obviously a lack of available bandwidth/QoS in either direction is going to cause it grief, but that has nothing to do with the ACK situation.

In this case having some upstream ACKs getting lost may even be beneficial to the video conf if it causes the big TCP download that is happening to slow down as a result of lost ACK's ;)
Mark Newton
2013-09-16 06:28:43 UTC
Permalink
Post by Tony
All video conf & VoIP I've seen uses UDP which means the video conferencing session to the medical professional is not going to care about ACK's at all.
Skype, Google Hangouts and Apple Facetime use TCP.

- mark
Tony
2013-09-16 11:05:20 UTC
Permalink
----- Original Message -----
Post by Tony
Post by Tony
All video conf & VoIP I've seen uses UDP which means the video
conferencing session to the medical professional is not going to care about
ACK's at all.
Skype, Google Hangouts and Apple Facetime use TCP.
Both Skype & Google attempt to use UDP first and only fall back to TCP if UDP is blocked/doesn't work. Most standard home/soho router/firewalls would I imagine allow the preferred UDP option to work. I tested a skype call just now and it used UDP and that works outbound through my firewall which is an old snapgear with a fairly default ruleset (couple of NAT ports inbound to servers, allow anything outbound).

Don't know about Facetime, the FAQ on ports it uses is unclear which is used for what purpose, but something I read suggests that it is much the same and basically uses SIP signalling (which uses TCP of course) with a UDP RTP stream that is wrapped around with some registration, authentication and other Apple stuff. It may also fall back to using TCP for transport of the RTP stream if UDP fails, this wasn't clear to me in anything I read.

I'd also be fairly dissapointed if my medical professional was going to use any of those three video conf solutions for a remote "consultation". Having said that I can see that they are cheap, widely spread, easy to use & easily available to the masses, so maybe they would ?

Aren't we all moving to IPv6 anyway at which point all of these three video conf solutions (Skype, Google, Apple) will be able to use their preferred UDP when NAT ugliness goes away with v6 ?!


regards,
Tony.


http://en.wikipedia.org/wiki/Skype_protocol#Login
http://support.google.com/a/bin/answer.py?hl=en&answer=1279090
http://support.apple.com/kb/HT4245
Matthew Moyle-Croft
2013-09-16 15:50:56 UTC
Permalink
Bullish about VDSL2 in Australia -

Loading Image...
Loading Image...

Enjoy.

MMC
Post by Mark ZZZ Smith
----- Original Message -----
Post by Tony
Post by Tony
All video conf & VoIP I've seen uses UDP which means the video
conferencing session to the medical professional is not going to care about
ACK's at all.
Skype, Google Hangouts and Apple Facetime use TCP.
Both Skype & Google attempt to use UDP first and only fall back to TCP if UDP is blocked/doesn't work. Most standard home/soho router/firewalls would I imagine allow the preferred UDP option to work. I tested a skype call just now and it used UDP and that works outbound through my firewall which is an old snapgear with a fairly default ruleset (couple of NAT ports inbound to servers, allow anything outbound).
Don't know about Facetime, the FAQ on ports it uses is unclear which is used for what purpose, but something I read suggests that it is much the same and basically uses SIP signalling (which uses TCP of course) with a UDP RTP stream that is wrapped around with some registration, authentication and other Apple stuff. It may also fall back to using TCP for transport of the RTP stream if UDP fails, this wasn't clear to me in anything I read.
I'd also be fairly dissapointed if my medical professional was going to use any of those three video conf solutions for a remote "consultation". Having said that I can see that they are cheap, widely spread, easy to use & easily available to the masses, so maybe they would ?
Aren't we all moving to IPv6 anyway at which point all of these three video conf solutions (Skype, Google, Apple) will be able to use their preferred UDP when NAT ugliness goes away with v6 ?!
regards,
Tony.
http://en.wikipedia.org/wiki/Skype_protocol#Login
http://support.google.com/a/bin/answer.py?hl=en&answer=1279090
http://support.apple.com/kb/HT4245
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Paul Brooks
2013-09-16 23:00:03 UTC
Permalink
Post by Matthew Moyle-Croft
Bullish about VDSL2 in Australia -
http://mmc.com.au/pub/DodgyPhoneLine2.jpg
http://mmc.com.au/pub/DodgyPhoneLine.jpg
Enjoy.
...and then see "Skin Effect" (http://www.allaboutcircuits.com/vol_2/chpt_3/6.html).
At 10 MHz, the Skin Effect depth is a channel of depth 0.02mm - that is, only a ring
of the outer 10% of the copper wire actually carries any signal at 10 MHz, even
thinner for higher frequencies.
Precisely where surface corrosion first starts to pit and erode the conductor.

Enjoy.

Joseph Goldman
2013-09-15 04:35:20 UTC
Permalink
For most RSP's who buy transit at a 1:1 ratio, unless they happen to
offset a lot of bandwidth with content hosting, then you tend to have a
lot of spare upload spare anyway. I wouldn't see the problem in
symmetric uploads being sold. Obviously each RSP can have T&C's to state
abuse and what can/can't be hosted at home (i.e. a big shared hosting
company can't be hosted off your 50/50 NBN Tail)
Post by Mark ZZZ Smith
----- Original Message -----
Sent: Saturday, 14 September 2013 11:22 AM
Subject: Re: [AusNOG] ADSL2+ line sync data
[..]
Post by Mark ZZZ Smith
It's the ratio of downstream to update bandwidth that matters, and
the likelihood of congestion in the upstream direction, not so much
the bandwidth involved. The greater the ratio of downstream to
upstream bandwidth, the more likely the problem is going to occur.
+1
(Seriously, Mark is right about http://tools.ietf.org/html/rfc3449. Still
relevant in our brave new asymmetric high speed world.)
Where I most think it matters is that the service speeds people buy sets, quite reasonably, their expectations of what they'll be able to get out of their service.
So if a customer buys a 25/5 service for example, they'd expect that they'd be able to upload at 5Mbps while also downloading at 25Mbps. They probably won't be able to due to bandwidth asymmetry, which means that the ISP/RSP may be violating trade practices laws for false advertising. Or in the least, the ISP/RSP have to deal with customer complaints.
OTOH, if the ISP/RSP provided 25/25, then it is possible to both upload and download using TCP at full rates in both directions.
In the brave new NBN world, where people have data centers in their basement, are uploading X-rays from their iPhone 9, and streaming 4K security camera video of their cat to the Internet, I think the consequences of this asymmetry are going to become are more visible.
Regards,
Mark.
cheers,
gja
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Mark ZZZ Smith
2013-09-15 05:32:30 UTC
Permalink
----- Original Message -----
Sent: Sunday, 15 September 2013 2:35 PM
Subject: Re: [AusNOG] ADSL2+ line sync data
For most RSP's who buy transit at a 1:1 ratio, unless they happen to
offset a lot of bandwidth with content hosting, then you tend to have a
lot of spare upload spare anyway. I wouldn't see the problem in
symmetric uploads being sold. Obviously each RSP can have T&C's to state
abuse and what can/can't be hosted at home (i.e. a big shared hosting
company can't be hosted off your 50/50 NBN Tail)
 
Assuming the existing NBN model/services, an RSP/ISP could buy 100/40 or 50/20 services from NBNco, but then on the BNG/BRAS, shape (not police) the downstream service bandwidth to be equal to the upstream bandwidth, so that their service bandwidths were 40/40 or 20/20. They'd then need less CVC and Internet transit capacity, and therefore should be able to offer these services significantly cheaper that their competitors 100/40 or 50/20 services (simplistically, the 40/40 service could be 60% cheaper than their competitors 100/40 service).
Post by Mark ZZZ Smith
----- Original Message -----
Sent: Saturday, 14 September 2013 11:22 AM
Subject: Re: [AusNOG] ADSL2+ line sync data
      [..]
  It's the ratio of downstream to update bandwidth that
matters, and
Post by Mark ZZZ Smith
  the likelihood of congestion in the upstream direction, not so
much
Post by Mark ZZZ Smith
  the bandwidth involved. The greater the ratio of downstream to
  upstream bandwidth, the more likely the problem is going to
occur.
Post by Mark ZZZ Smith
+1
(Seriously, Mark is right about http://tools.ietf.org/html/rfc3449.
Still
Post by Mark ZZZ Smith
relevant in our brave new asymmetric high speed world.)
 
Where I most think it matters is that the service speeds people buy sets,
quite reasonably, their expectations of what they'll be able to get out of
their service.
Post by Mark ZZZ Smith
So if a customer buys a 25/5 service for example, they'd expect that
they'd be able to upload at 5Mbps while also downloading at 25Mbps. They
probably won't be able to due to bandwidth asymmetry, which means that the
ISP/RSP may be violating trade practices laws for false advertising. Or in the
least, the ISP/RSP have to deal with customer complaints.
Post by Mark ZZZ Smith
OTOH, if the ISP/RSP provided 25/25, then it is possible to both upload and
download using TCP at full rates in both directions.
Post by Mark ZZZ Smith
In the brave new NBN world, where people have data centers in their
basement, are uploading X-rays from their iPhone 9, and streaming 4K security
camera video of their cat to the Internet, I think the consequences of this
asymmetry are going to become are more visible.
Post by Mark ZZZ Smith
Regards,
Mark.
cheers,
gja
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Joshua D'Alton
2013-09-15 05:58:18 UTC
Permalink
Could be 60% cheaper, but wouldn't be. Given they'd be using a
quota/allowance for the service, selling 40/40 vs 100/40 would only impact
the amount of burstability and therefore the possibility of the ISPs
transit being saturated, but not by much.

I'm not assuming residential users only download, just that the majority is
download, and that were it to come to a decision of 50/20 vs 40/40
(symetrical from a provider doing said shaping), 50/20 will win just about
every time.

I think you're missing that I'm not disputing that symmetry isn't
beneficial, or that in practice it won't be for some people, because it is
and it will, but that most of the time they *won't* benefit, and indeed
most of the time they'd actually be experiencing detriment as a result of
having a slow symetrical service.


On Sun, Sep 15, 2013 at 3:32 PM, Mark ZZZ Smith
Post by Mark ZZZ Smith
----- Original Message -----
Sent: Sunday, 15 September 2013 2:35 PM
Subject: Re: [AusNOG] ADSL2+ line sync data
For most RSP's who buy transit at a 1:1 ratio, unless they happen to
offset a lot of bandwidth with content hosting, then you tend to have a
lot of spare upload spare anyway. I wouldn't see the problem in
symmetric uploads being sold. Obviously each RSP can have T&C's to state
abuse and what can/can't be hosted at home (i.e. a big shared hosting
company can't be hosted off your 50/50 NBN Tail)
Assuming the existing NBN model/services, an RSP/ISP could buy 100/40 or
50/20 services from NBNco, but then on the BNG/BRAS, shape (not police) the
downstream service bandwidth to be equal to the upstream bandwidth, so that
their service bandwidths were 40/40 or 20/20. They'd then need less CVC and
Internet transit capacity, and therefore should be able to offer these
services significantly cheaper that their competitors 100/40 or 50/20
services (simplistically, the 40/40 service could be 60% cheaper than their
competitors 100/40 service).
Post by Mark ZZZ Smith
----- Original Message -----
Sent: Saturday, 14 September 2013 11:22 AM
Subject: Re: [AusNOG] ADSL2+ line sync data
[..]
Post by Mark ZZZ Smith
It's the ratio of downstream to update bandwidth that
matters, and
Post by Mark ZZZ Smith
Post by Mark ZZZ Smith
the likelihood of congestion in the upstream direction, not so
much
Post by Mark ZZZ Smith
Post by Mark ZZZ Smith
the bandwidth involved. The greater the ratio of downstream to
upstream bandwidth, the more likely the problem is going to
occur.
Post by Mark ZZZ Smith
+1
(Seriously, Mark is right about http://tools.ietf.org/html/rfc3449.
Still
Post by Mark ZZZ Smith
relevant in our brave new asymmetric high speed world.)
Where I most think it matters is that the service speeds people buy
sets,
quite reasonably, their expectations of what they'll be able to get out
of
their service.
Post by Mark ZZZ Smith
So if a customer buys a 25/5 service for example, they'd expect that
they'd be able to upload at 5Mbps while also downloading at 25Mbps. They
probably won't be able to due to bandwidth asymmetry, which means that
the
ISP/RSP may be violating trade practices laws for false advertising. Or
in the
least, the ISP/RSP have to deal with customer complaints.
Post by Mark ZZZ Smith
OTOH, if the ISP/RSP provided 25/25, then it is possible to both
upload and
download using TCP at full rates in both directions.
Post by Mark ZZZ Smith
In the brave new NBN world, where people have data centers in their
basement, are uploading X-rays from their iPhone 9, and streaming 4K
security
camera video of their cat to the Internet, I think the consequences of
this
asymmetry are going to become are more visible.
Post by Mark ZZZ Smith
Regards,
Mark.
cheers,
gja
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
_______________________________________________
AusNOG mailing list
http://lists.ausnog.net/mailman/listinfo/ausnog
Rod Veith
2013-09-16 05:20:39 UTC
Permalink
Seriously guys, we've got stop equating maximum line speeds to data transfer
rates (eg download speeds). They are 2 completely different measurements. It
is no wonder customers get confused when people in the industry, and who
should know better, keep saying that download speed = line speed. They are
NOT.

Rod



So if a customer buys a 25/5 service for example, they'd expect that they'd
be able to upload at 5Mbps while also downloading at 25Mbps. They probably
won't be able to due to bandwidth asymmetry, which means that the ISP/RSP
may be violating trade practices laws for false advertising. Or in the
least, the ISP/RSP have to deal with customer complaints.

OTOH, if the ISP/RSP provided 25/25, then it is possible to both upload and
download using TCP at full rates in both directions.

In the brave new NBN world, where people have data centers in their
basement, are uploading X-rays from their iPhone 9, and streaming 4K
security camera video of their cat to the Internet, I think the consequences
of this asymmetry are going to become are more visible.


Regards,

Mark.
Mark ZZZ Smith
2013-09-14 01:31:59 UTC
Permalink
----- Original Message -----
Sent: Saturday, 14 September 2013 10:15 AM
Subject: Re: [AusNOG] ADSL2+ line sync data
Post by Paul Brooks
________________________________
Sent: Friday, 13 September 2013 10:25 PM
Subject: Re: [AusNOG] ADSL2+ line sync data
<snip>
Symmetry doesn't eliminate congestion (nothing does), 
And to clarify what I mean by this, TCP attempts to push the network into congestion to measure and then maximise the use of the available capacity between the end points. When packets are lost, TCP considers that to be a sign of congestion and backs off its send rate.
but it can provide
equal performance and quality in both directions for TCP and other protocols and
applications. Video can be the same quality in both directions, not better in
one direction than the other (asymmetry of bandwidth is going to put a real
cramp on telemedicine, as the patient will see the doctor really well, but the
doctor won't see the patient really well - and doctors won't accept that
because their malpractice premiums would then go up). Uploads will have as much
congestion effect as downloads. 
Post by Paul Brooks
I don't know exactly what ratio is 'ideal',
True, but the IETF do, and they're the experts on Internet protocols. They
designed TCP, and they've published an RFC that has become a Best Common
Practice on how asymmetry impacts its performance.
If you read nothing else, read the following from the summary,
"Asymmetry, and particular high asymmetry, raises a set of TCP performance issues."
Joshua D'Alton
2013-09-14 03:25:51 UTC
Permalink
Yea perhaps I should have said "isn't as applicable for home connections".
All your rebuttal points I agree with completely, I guess the point I was
making was, that in terms of "raises a set of TCP performance issues", --->
it does, but when it comes to home xDSL FTTx that the theory might not be
as relevant when it comes to the actual reality of the user use cases.

As in "Sad thing about this is that doing that appears to be a good idea,
but that actual best thing for performance is symmetry", true in theory and
some realities, but in home connectivity reality 18/1 will beat 11/11 the
vast majority of the time and therefore the users will pick that the
majority of the time. The cases where they don't being the cases they
actually need sufficient upload to negate the small difference in download
potential.

Anyway, interesting subject I'll take my foot out of my mouth now!


On Sat, Sep 14, 2013 at 11:31 AM, Mark ZZZ Smith
Post by Mark ZZZ Smith
----- Original Message -----
Sent: Saturday, 14 September 2013 10:15 AM
Subject: Re: [AusNOG] ADSL2+ line sync data
Post by Paul Brooks
________________________________
Sent: Friday, 13 September 2013 10:25 PM
Subject: Re: [AusNOG] ADSL2+ line sync data
<snip>
Symmetry doesn't eliminate congestion (nothing does),
And to clarify what I mean by this, TCP attempts to push the network into
congestion to measure and then maximise the use of the available capacity
between the end points. When packets are lost, TCP considers that to be a
sign of congestion and backs off its send rate.
but it can provide
equal performance and quality in both directions for TCP and other
protocols and
applications. Video can be the same quality in both directions, not
better in
one direction than the other (asymmetry of bandwidth is going to put a
real
cramp on telemedicine, as the patient will see the doctor really well,
but the
doctor won't see the patient really well - and doctors won't accept that
because their malpractice premiums would then go up). Uploads will have
as much
congestion effect as downloads.
Post by Paul Brooks
I don't know exactly what ratio is 'ideal',
True, but the IETF do, and they're the experts on Internet protocols.
They
designed TCP, and they've published an RFC that has become a Best Common
Practice on how asymmetry impacts its performance.
If you read nothing else, read the following from the summary,
"Asymmetry, and particular high asymmetry, raises a set of TCP
performance
issues."
Peter Tonoli
2013-09-16 06:06:15 UTC
Permalink
Hi,
Post by Tony
All video conf & VoIP I've seen uses UDP which means the video
conferencing session to the medical professional is not going to care
about ACK's at all. Obviously a lack of available bandwidth/QoS in
either direction is going to cause it grief, but that has nothing to
do with the ACK situation.
Skype uses UDP & TCP for multimedia payload and TCP for signallying - especially TCP when trying to break through restrictive firewalls.

P.
--
Peter Tonoli < ***@medstv.unimelb.edu.au > +61-3-9288-2399
IT Manager
The University of Melbourne - Eastern Hill Academic Centre, St. Vincent's Institute and O'Brien Institute
Continue reading on narkive:
Loading...