Simon Hackett dispels the myths, promoted from a 2006 Citigroup report, that claim ADSL2+ has poor performance in the real world. Simon, the managing director of national broadband trailblazer Internode and network infrastructructure pioneer Agile Communications, first published this article about the history of ADSL broadband in Australia on his blog www.simonhackett.com.

Internode was the first company to launch commercial ADSL2+ services in Australia well over a year ago.

As such, Internode, along with other organisations such as iiNet that were early movers in deploying ADSL2+ services, has lot of real world experience with the achievable and sustainable data speeds available in Australia using this technology. This is experience that Telstra does not have.

Various members of the media have, this week, cited a report by Citigroup about FTTN and ADSL2+ which is, simply, codswallop.

To my chagrin, many respectable journalist have cited and repeated these most remarkably wrong claims, as if they were fact.

These statements, which claim ADSL2+ can only deliver a realistic speed of about 3 megabits per second (Mbps) on copper lines averaging distances of 2km are simply rubbish.

Accuracy about real world ADSL2+ performance is critical to the entire debate over this issue. Inaccurate claims only perpetuate the Telstra party line that its mysterious and never fully revealed FTTN proposals held a secret magic pudding that somehow works better than ADSL2+, and that is somehow different in some way to existing FTTN deployments.

The truth is that speeds touted by Telstra as requiring FTTN are already available, and are being enjoyed today, across Australia.

This is demonstrated by the document at the following URL which contains highly relevant real-world data:

http://www.iinet.net.au/about/media/releases/The-Myth-of-Fibre-May-06.pdf

This paper from iiNet provides actual data points from the iiNet ADSL2+ deployment in the Sydney basin. These are real world Australian results, not consultants projections or assumptions based on ... well, certainly not based on fact.

Internode’s experience with its ADSL2+ deployment mirrors that of iiNet in this regard.

More than 90 per cent of our ADSL2+ customers in metro areas enjoy speeds faster than 6Mb/s today.

With Internode’s impending launch of ADSL2+ Annex M services â?? which are already used on a trial basis by a number of Internode customers â?? we are now adding to our capabilities by being able to offer upstream speeds of the order of 2 Mbps.

This upstream data rate is well beyond the paltry 256 kilobit upload data rate that Telstra has locked most of its own ADSL its lines to since 1999.

If you look at the Internode Speed v. Distance graph at http://www.internode.on.net/adsl2/graph/index.htm, you can see the dramatic speed advantage of unconstrained ADSL compared with the anaemic 1.5Mb/s limit that Telstra has administratively enforced in its network since 1999.

Even Australia’s first generation ADSL is capable of data transfer speeds far above the numbers cited from the Citigroup report at distances far beyond the 1.5km line length that Telstra has promoted as a drop-dead limit for high performance broadband over copper.

The 1.5km limitation assumed this “magic” limitation status because Telstra propagated this limit during the farcical FTTN debate and various parts of the media picked it up and copied it - regardless of the demonstrated reality in Australia.

So, lets specifically note and then rebut every bullet point ascribed to that Citigroup report by various Australian journalists this week.

These bullet points are the claimed reasons why ADSL2+ just doesn’t work well in Australia, and (implicitly) why only ADSL2+ with a Telstra badge on the cabinet would work better:

“The thin diameter of Australian copper"

Australian copper is not unusually thin at all. Indeed, typically the opposite is the case.

Internode’s experience with rural ADSL2+ deployments demonstrate that the long distances covered by copper in country areas (and the corresponding choice of heavier gauge copper in those runs) mean that it is routine for our customers in regional zones to achieve speeds above 6Mb/s at distances far exceeding those in the city.

“For lengths longer than 1.5km, ADSL2+ behaves like ADSL”

The graph at http://www.internode.on.net/adsl2/graph/index.htm displays that the performance achieved by ADSL2+ remains far in excess of that achievable by ADSL.

Indeed, even ADSL1 is operating much faster than 1.5Mb/s far beyond that distance. Telstra has made a decision to keep its peak network speed in 2006 â?? six years down the track â?? to 1.5Mb/s as per the red Telstra administrative constraint box on that graph.

It is important to appreciate that ADSL1 â?? the standard deployed in Australia since 1999 â?? was capable of speeds of up to 8Mb/s on day one. Constraint below that speed is a business decision of Telstra, not a technical limitation. Recently, it has been touted as due to technical limitations to advance the underlying deceit in the selling of FTTN to the market. Put simply, it ain't so.

“Cross-talk problems”

Internode has practical experience which shows that while cross-talk does degrade performance to a mild extent, as a designed-in part of the protocol, it’s entirely manageable in practice. Years of careful deployment rule design by ACIF, with Telstra as a key participant, have ensured that the mangement of crosstalk in the copper line network is entirely tenable.

“Long submarine links to content in the US and Europe”

Long cables don't slow data flow down, thin or full ones do.

Internode operates at this time (August 2006) almost three gigabits per second of leased capacity on the Southern Cross Cable Network in its own right. Our customers hit the peak performance of their ADSL2+ Extreme broadband services at any time they wish to â?? 24x7x365.

Properly engineered international networks work fine – and, by the way, are absolutely unrelated to the engineering of properly deployed local loop networks. A great ISP engineers its whole network well. A lacklustre one does not. This is nothing to do with ADSL2+ and everything to do with professional deployments v. “cowboy” ones.

“Filters present in all self-install kits, used by 85 per cent of homes”

These filters work just fine. The latest generation of filters comply with all of the applicable standards and do what they're supposed to do.

Internode customers see fast, real-world speeds using our supplied filters.

Occasional cases of problems with internal house wiring (or a need to interwork properly with back-to-base alarms) can be handled with the (routine and inexpensive) installation of a ‘Central Splitter” into the premises by a licensed cabler.

“Lack of backhaul capacity from the exchange”

This may be a problem for Telstra, in its 1999 generation ATM backhaul network, but its an utter non-issue for its competitors, such as Internode and iiNet. We use 'dark fibre' â?? i.e. fibre operating at an unconstrained data rate â?? leased from dark fibre service providers such as Pipe Networks. Those links operate at multi-gigabit speeds, and are expandable by changing the equipment at each end, without any further expenditure on the access costs to use those fibres.

Non-Telstra ADSL2+ deployments are already built in this 'next generation network' manner that Telstra is scrambling to copy. Backhaul is not a problem for any properly engineered competitor's network. Indeed, there are already plenty of ADSL2+ competitors for Telstra â?? in exactly the same coverage areas that FTTN was expected to serve.

In conclusion, media coverage of the ‘facts’ about ADSL2+ lately contain a combination of falsehoods and misunderstandings as well as descriptions of constraints that exist on the Telstra network but not in any elses’ network.

In the real world, Telstra’s competitors â?? who have sold ADSL2+ services for more than a year, in essentially the same coverage area that the FTTN farce would have covered â?? are providing customers with the services that Telstra claimed to need AUS$4bn and the destruction of a decade of competition reform (by physically unplugging 100% of their competitors, permanently) to achieve.

Telstra’s FTTN campaign has worked to pull the wool over the eyes of its shareholders, the public and the media for more than a year.

Proper scrutiny produces the revelation that not only does the emperor have no clothes, but the competitors to the emperor have been selling their customers a fine choice of designer clothing since March 2006.

The truth is that the broadband situation in Australia is not bad. It’s actually great - and getting better by the week - without FTTN.

There is a strong and exciting scope, now that FTTN is dead, for innovative competitors, and for groups like the G9 to take the ball and run with it - to deploy even more cost effective, and more efficient, ways to improve broadband outcomes in Australia.

The paying customers of Internode (and other companies deploying ADSL2+ in Australia) are providing, daily, that ‘real’ broadband is extremely available in Australia - today - and its deployment is becoming more and more widespread by the week.

Well within the touted build timeframe for ‘FTTN’ (’40 months’), more than 90% of customers who want ‘faster than 1.5Mb/s’ broadband in Australia will be able to buy it, from multiple service providers.

The remaining 10% can be addressed by the efforts of the G9 and other individual service providers in various ways. This can be accelerated if Telstra decide to become a part of the solution, but one way or another - we’ll get there.

Twitter